4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1 [xx XXX xxxx]
27 * Add a mac salt length option for the pkcs12 command.
31 * Add more SRTP protection profiles from RFC8723 and RFC8269.
35 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
37 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
39 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
40 supported and enabled.
44 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
45 to the list of ciphersuites providing Perfect Forward Secrecy as
46 required by SECLEVEL >= 3.
48 *Dmitry Belyavskiy, Nicola Tuveri*
50 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting. The
51 SSL_CTRL_GET_IANA_GROUPS control code, exposed as the SSL_get0_iana_groups()
52 function-like macro, retrieves the list of supported groups sent by the peer,
53 and the function SSL_client_hello_get_extension_order() populates a caller-supplied
54 array with the list of extension types present in the ClientHello, in order of
59 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
60 to make it possible to use empty passphrase strings.
64 * RNDR and RNDRRS support in provider functions to provide
65 random number generation for Arm CPUs (aarch64).
69 * s_client and s_server apps now explicitly say when the TLS version
70 does not include the renegotiation mechanism. This avoids confusion
71 between that scenario versus when the TLS version includes secure
72 renegotiation but the peer lacks support for it.
76 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
78 *Tomasz Kantecki, Andrey Matyukov*
80 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
81 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
82 of 160 bits and above and less than 224 bits were previously accepted by
83 default but are now no longer allowed. By default TLS compression was
84 already disabled in previous OpenSSL versions. At security level 2 it cannot
89 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
94 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
95 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
96 will need to load the legacy crypto provider.
100 * The various OBJ_* functions have been made thread safe.
104 * CCM8 cipher suites in TLS have been downgraded to security level zero
105 because they use a short authentication tag which lowers their strength.
109 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
114 * Parallel dual-prime 1536/2048-bit modular exponentiation for
115 AVX512_IFMA capable processors.
117 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
119 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
120 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
121 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
122 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
123 `OPENSSL_NO_DEPRECATED_3_1`.
125 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
126 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
127 definitions for these functions regardless of whether
128 `OPENSSL_NO_DEPRECATED_3_1` is defined.
130 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
131 functions regardless of whether they are using them. It is recommended that
132 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
139 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
140 listed here are only a brief description.
141 The migration guide contains more detailed information related to new features,
142 breaking changes, and mappings for the large list of deprecated functions.
144 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
146 ### Changes between 3.0.3 and 3.0.4 [21 June 2022]
148 * In addition to the c_rehash shell command injection identified in
149 CVE-2022-1292, further bugs where the c_rehash script does not
150 properly sanitise shell metacharacters to prevent command injection have been
153 When the CVE-2022-1292 was fixed it was not discovered that there
154 are other places in the script where the file names of certificates
155 being hashed were possibly passed to a command executed through the shell.
157 This script is distributed by some operating systems in a manner where
158 it is automatically executed. On such operating systems, an attacker
159 could execute arbitrary commands with the privileges of the script.
161 Use of the c_rehash script is considered obsolete and should be replaced
162 by the OpenSSL rehash command line tool.
165 *Daniel Fiala, Tomáš Mráz*
167 * Case insensitive string comparison no longer uses locales. It has instead
168 been directly implemented.
172 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
174 * Case insensitive string comparison is reimplemented via new locale-agnostic
175 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
176 comparison. The previous implementation had problems when the Turkish locale
181 * Fixed a bug in the c_rehash script which was not properly sanitising shell
182 metacharacters to prevent command injection. This script is distributed by
183 some operating systems in a manner where it is automatically executed. On
184 such operating systems, an attacker could execute arbitrary commands with the
185 privileges of the script.
187 Use of the c_rehash script is considered obsolete and should be replaced
188 by the OpenSSL rehash command line tool.
193 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
194 certificate on an OCSP response. The bug caused the function in the case
195 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
196 response (meaning a successful verification) even in the case where the
197 response signing certificate fails to verify.
199 It is anticipated that most users of `OCSP_basic_verify` will not use the
200 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
201 a negative value (indicating a fatal error) in the case of a certificate
202 verification failure. The normal expected return value in this case would be
205 This issue also impacts the command line OpenSSL "ocsp" application. When
206 verifying an ocsp response with the "-no_cert_checks" option the command line
207 application will report that the verification is successful even though it
208 has in fact failed. In this case the incorrect successful response will also
209 be accompanied by error messages showing the failure and contradicting the
210 apparently successful result.
215 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
216 AAD data as the MAC key. This made the MAC key trivially predictable.
218 An attacker could exploit this issue by performing a man-in-the-middle attack
219 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
220 that the modified data would still pass the MAC integrity check.
222 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
223 endpoint will always be rejected by the recipient and the connection will
224 fail at that point. Many application protocols require data to be sent from
225 the client to the server first. Therefore, in such a case, only an OpenSSL
226 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
228 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
229 sent in both directions. In this case both clients and servers could be
230 affected, regardless of the application protocol.
232 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
233 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
234 the handshake when using this ciphersuite.
236 The confidentiality of data is not impacted by this issue, i.e. an attacker
237 cannot decrypt data that has been encrypted using this ciphersuite - they can
240 In order for this attack to work both endpoints must legitimately negotiate
241 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
242 OpenSSL 3.0, and is not available within the default provider or the default
243 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
244 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
245 following must have occurred:
247 1) OpenSSL must have been compiled with the (non-default) compile time option
248 enable-weak-ssl-ciphers
250 2) OpenSSL must have had the legacy provider explicitly loaded (either
251 through application code or via configuration)
253 3) The ciphersuite must have been explicitly added to the ciphersuite list
255 4) The libssl security level must have been set to 0 (default is 1)
257 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
259 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
260 others that both endpoints have in common
265 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
266 occuppied by the removed hash table entries.
268 This function is used when decoding certificates or keys. If a long lived
269 process periodically decodes certificates or keys its memory usage will
270 expand without bounds and the process might be terminated by the operating
271 system causing a denial of service. Also traversing the empty hash table
272 entries will take increasingly more time.
274 Typically such long lived processes might be TLS clients or TLS servers
275 configured to accept client certificate authentication.
278 *Hugo Landau, Aliaksei Levin*
280 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
281 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
282 statistics are no longer supported. For compatibility, these statistics are
283 still listed in the output but are now always reported as zero.
287 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
289 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
290 for non-prime moduli.
292 Internally this function is used when parsing certificates that contain
293 elliptic curve public keys in compressed form or explicit elliptic curve
294 parameters with a base point encoded in compressed form.
296 It is possible to trigger the infinite loop by crafting a certificate that
297 has invalid explicit curve parameters.
299 Since certificate parsing happens prior to verification of the certificate
300 signature, any process that parses an externally supplied certificate may thus
301 be subject to a denial of service attack. The infinite loop can also be
302 reached when parsing crafted private keys as they can contain explicit
303 elliptic curve parameters.
305 Thus vulnerable situations include:
307 - TLS clients consuming server certificates
308 - TLS servers consuming client certificates
309 - Hosting providers taking certificates or private keys from customers
310 - Certificate authorities parsing certification requests from subscribers
311 - Anything else which parses ASN.1 elliptic curve parameters
313 Also any other applications that use the BN_mod_sqrt() where the attacker
314 can control the parameter values are vulnerable to this DoS issue.
319 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
320 to the list of ciphersuites providing Perfect Forward Secrecy as
321 required by SECLEVEL >= 3.
323 *Dmitry Belyavskiy, Nicola Tuveri*
325 * Made the AES constant time code for no-asm configurations
326 optional due to the resulting 95% performance degradation.
327 The AES constant time code can be enabled, for no assembly
328 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
332 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
337 * The negative return value handling of the certificate verification callback
338 was reverted. The replacement is to set the verification retry state with
339 the SSL_set_retry_verify() function.
343 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
345 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
346 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
347 verify a certificate supplied by a server. That function may return a
348 negative return value to indicate an internal error (for example out of
349 memory). Such a negative return value is mishandled by OpenSSL and will cause
350 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
351 success and a subsequent call to SSL_get_error() to return the value
352 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
353 returned by OpenSSL if the application has previously called
354 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
355 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
356 totally unexpected and applications may not behave correctly as a result. The
357 exact behaviour will depend on the application but it could result in
358 crashes, infinite loops or other similar incorrect responses.
360 This issue is made more serious in combination with a separate bug in OpenSSL
361 3.0 that will cause X509_verify_cert() to indicate an internal error when
362 processing a certificate chain. This will occur where a certificate does not
363 include the Subject Alternative Name extension but where a Certificate
364 Authority has enforced name constraints. This issue can occur even with valid
370 * Corrected a few file name and file reference bugs in the build,
371 installation and setup scripts, which lead to installation verification
372 failures. Slightly enhanced the installation verification script.
376 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
381 * Fixed PVK encoder to properly query for the passphrase.
385 * Multiple fixes in the OSSL_HTTP API functions.
389 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
390 OSSL_PARAM_INTEGER data type and return error on negative numbers
391 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
392 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
396 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
400 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
404 * Multiple threading fixes.
408 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
412 * Allow fetching an operation from the provider that owns an unexportable key
413 as a fallback if that is still allowed by the property query.
417 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
419 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
424 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
425 S390X capability vector to zero. This simplifies testing of different code
426 paths on S390X architecture.
430 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
431 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
432 SP 800-38D". The communication will fail at this point.
436 * The EC_GROUP_clear_free() function is deprecated as there is nothing
437 confidential in EC_GROUP data.
441 * The byte order mark (BOM) character is ignored if encountered at the
442 beginning of a PEM-formatted file.
446 * Added CMS support for the Russian GOST algorithms.
450 * Due to move of the implementation of cryptographic operations
451 to the providers, validation of various operation parameters can
452 be postponed until the actual operation is executed where previously
453 it happened immediately when an operation parameter was set.
455 For example when setting an unsupported curve with
456 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
457 fail but later keygen operations with the EVP_PKEY_CTX will fail.
459 *OpenSSL team members and many third party contributors*
461 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
462 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
463 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
464 instead to retrieve these algorithms from a provider.
468 * On build targets where the multilib postfix is set in the build
469 configuration the libdir directory was changing based on whether
470 the lib directory with the multilib postfix exists on the system
471 or not. This unpredictable behavior was removed and eventual
472 multilib postfix is now always added to the default libdir. Use
473 `--libdir=lib` to override the libdir if adding the postfix is
478 * The triple DES key wrap functionality now conforms to RFC 3217 but is
479 no longer interoperable with OpenSSL 1.1.1.
483 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
484 function codes, this function can only cause problems for calling
489 * Add a configurable flag to output date formats as ISO 8601. Does not
490 change the default date format.
494 * Version of MSVC earlier than 1300 could get link warnings, which could
495 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
496 Support for this flag has been removed.
500 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
501 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
502 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
503 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
504 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
508 * The signatures of the functions to get and set options on SSL and
509 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
510 Some source code changes may be required.
514 * The public definitions of conf_method_st and conf_st have been
515 deprecated. They will be made opaque in a future release.
517 *Rich Salz and Tomáš Mráz*
519 * Client-initiated renegotiation is disabled by default. To allow it, use
520 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
521 flag, or the "ClientRenegotiation" config parameter as appropriate.
525 * Add "abspath" and "includedir" pragma's to config files, to prevent,
526 or modify relative pathname inclusion.
530 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
531 validated. Please consult the README-FIPS and
532 README-PROVIDERS files, as well as the migration guide.
534 *OpenSSL team members and many third party contributors*
536 * For the key types DH and DHX the allowed settable parameters are now different.
540 * The openssl commands that read keys, certificates, and CRLs now
541 automatically detect the PEM or DER format of the input files.
543 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
545 * Added enhanced PKCS#12 APIs which accept a library context.
549 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
553 * Added support for Kernel TLS (KTLS).
555 *Boris Pismenny, John Baldwin and Andrew Gallatin*
557 * Support for RFC 5746 secure renegotiation is now required by default for
558 SSL or TLS connections to succeed.
562 * The signature of the `copy` functional parameter of the
563 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
564 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
565 the signature of the `pub_decode` functional parameter of the
566 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
567 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
571 * The error return values from some control calls (ctrl) have changed.
575 * A public key check is now performed during EVP_PKEY_derive_set_peer().
579 * Many functions in the EVP_ namespace that are getters of values from
580 implementations or contexts were renamed to include get or get0 in their
581 names. Old names are provided as macro aliases for compatibility and
586 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
587 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
588 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
593 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
596 * The output from the command line applications may have minor
601 * The output from numerous "printing" may have minor changes.
605 * Windows thread synchronization uses read/write primitives (SRWLock) when
606 supported by the OS, otherwise CriticalSection continues to be used.
610 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
611 work on read only BIO source/sinks that do not support these functions.
612 This allows piping or redirection of a file BIO using stdin to be buffered
613 into memory. This is used internally in OSSL_DECODER_from_bio().
617 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
618 this function would return one of the values OSSL_STORE_INFO_NAME,
619 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
620 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
621 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
622 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
623 using this function should be amended to handle the changed return value.
627 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
628 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
629 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
630 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
631 contains more than one certificate identifier: This means that all
632 certificates referenced there MUST be part of the validation chain.
636 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
637 RC5, DESX and DES have been moved to the legacy provider.
641 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
642 RIPEMD-160 have been moved to the legacy provider.
646 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
651 * The deprecated functions EVP_PKEY_get0_RSA(),
652 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
653 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
654 well as the similarly named "get1" functions behave differently in
659 * A number of functions handling low-level keys or engines were deprecated
660 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
661 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
662 EVP_PKEY_get0_siphash().
666 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
667 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
668 will need to load the legacy crypto provider. This includes these PBE
669 algorithms which use this KDF:
670 - NID_pbeWithMD2AndDES_CBC
671 - NID_pbeWithMD5AndDES_CBC
672 - NID_pbeWithSHA1AndRC2_CBC
673 - NID_pbeWithMD2AndRC2_CBC
674 - NID_pbeWithMD5AndRC2_CBC
675 - NID_pbeWithSHA1AndDES_CBC
679 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
680 BIO_debug_callback() functions.
684 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
685 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
689 * The RAND_METHOD APIs have been deprecated.
693 * The SRP APIs have been deprecated.
697 * Add a compile time option to prevent the caching of provider fetched
698 algorithms. This is enabled by including the no-cached-fetch option
699 at configuration time.
703 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
704 count of PKCS12_DEFAULT_ITER.
706 *Tomáš Mráz and Sahana Prasad*
708 * The openssl speed command does not use low-level API calls anymore.
712 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
715 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
717 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
721 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
722 providers may supply their own group implementations (using either the "key
723 exchange" or the "key encapsulation" methods) which will automatically be
724 detected and used by libssl.
726 *Matt Caswell, Nicola Tuveri*
728 * The undocumented function X509_certificate_type() has been deprecated;
732 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
736 * Removed RSA padding mode for SSLv23 (which was only used for
737 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
738 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
743 * Deprecated the obsolete X9.31 RSA key generation related functions.
745 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
746 is not allowed to return a value > 1, this is no more taken as failure.
748 *Viktor Dukhovni and David von Oheimb*
750 * Deprecated the obsolete X9.31 RSA key generation related functions
751 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
752 BN_X931_generate_prime_ex().
756 * The default key generation method for the regular 2-prime RSA keys was
757 changed to the FIPS 186-4 B.3.6 method.
761 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
765 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
769 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
770 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
772 *Rich Salz, Richard Levitte, and David von Oheimb*
774 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
778 * Deprecated `OCSP_parse_url()`.
782 * Validation of SM2 keys has been separated from the validation of regular EC
787 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
788 switches: a validation failure triggers an early exit, returning a failure
789 exit status to the parent process.
793 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
794 to ignore unknown ciphers.
798 * The `-cipher-commands` and `-digest-commands` options
799 of the command line utility `list` have been deprecated.
800 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
804 * Added convenience functions for generating asymmetric key pairs:
805 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
806 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
810 * All of the low level EC_KEY functions have been deprecated.
812 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
814 * Deprecated all the libcrypto and libssl error string loading
819 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
820 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
825 * The `-crypt` option to the `passwd` command line tool has been removed.
829 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
834 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
838 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
839 EVP_PKEY_get1_tls_encodedpoint().
843 * The security callback, which can be customised by application code, supports
844 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
845 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
849 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
850 interface. Their functionality remains unchanged.
854 * Added new option for 'openssl list', '-providers', which will display the
855 list of loaded providers, their names, version and status. It optionally
856 displays their gettable parameters.
860 * Removed EVP_PKEY_set_alias_type().
864 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
865 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
869 * Changed all "STACK" functions to be macros instead of inline functions. Macro
870 parameters are still checked for type safety at compile time via helper
875 * Remove the RAND_DRBG API
877 *Paul Dale and Matthias St. Pierre*
879 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
880 as well as actual hostnames.
884 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
885 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
886 conversely, silently ignore DTLS protocol version bounds when configuring
887 TLS-based contexts. The commands can be repeated to set bounds of both
888 types. The same applies with the corresponding "min_protocol" and
889 "max_protocol" command-line switches, in case some application uses both TLS
892 SSL_CTX instances that are created for a fixed protocol version (e.g.
893 `TLSv1_server_method()`) also silently ignore version bounds. Previously
894 attempts to apply bounds to these protocol versions would result in an
895 error. Now only the "version-flexible" SSL_CTX instances are subject to
896 limits in configuration files in command-line options.
900 * Deprecated the `ENGINE` API. Engines should be replaced with providers
905 * Reworked the recorded ERR codes to make better space for system errors.
906 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
907 given code is a system error (true) or an OpenSSL error (false).
911 * Reworked the test perl framework to better allow parallel testing.
913 *Nicola Tuveri and David von Oheimb*
915 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
916 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
920 * 'Configure' has been changed to figure out the configuration target if
921 none is given on the command line. Consequently, the 'config' script is
922 now only a mere wrapper. All documentation is changed to only mention
925 *Rich Salz and Richard Levitte*
927 * Added a library context `OSSL_LIB_CTX` that applications as well as
928 other libraries can use to form a separate context within which
929 libcrypto operations are performed.
933 * Added various `_ex` functions to the OpenSSL API that support using
934 a non-default `OSSL_LIB_CTX`.
938 * Handshake now fails if Extended Master Secret extension is dropped
943 * Dropped interactive mode from the `openssl` program.
947 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
949 *David von Oheimb and Shane Lontis*
951 * Deprecated `EC_METHOD_get_field_type()`.
955 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
956 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
957 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
961 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
965 * Add CAdES-BES signature verification support, mostly derived
966 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
968 *Filipe Raimundo da Silva*
970 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
974 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
975 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
979 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
983 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
984 EC_KEY_precompute_mult().
988 * Deprecated EC_POINTs_mul().
992 * Removed FIPS_mode() and FIPS_mode_set().
996 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1000 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1001 EC_POINT_get_Jprojective_coordinates_GFp().
1005 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1006 arrays to be more easily constructed via a series of utility functions.
1007 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1008 the various push functions and finally convert to a passable OSSL_PARAM
1009 array using OSSL_PARAM_BLD_to_param().
1013 * The security strength of SHA1 and MD5 based signatures in TLS has been
1018 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1019 contain a provider side internal key.
1023 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1027 * Project text documents not yet having a proper file name extension
1028 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1029 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1033 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1034 have been converted to Markdown with the goal to produce documents
1035 which not only look pretty when viewed online in the browser, but
1036 remain well readable inside a plain text editor.
1038 To achieve this goal, a 'minimalistic' Markdown style has been applied
1039 which avoids formatting elements that interfere too much with the
1040 reading flow in the text file. For example, it
1042 * avoids [ATX headings][] and uses [setext headings][] instead
1043 (which works for `<h1>` and `<h2>` headings only).
1044 * avoids [inline links][] and uses [reference links][] instead.
1045 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1047 [ATX headings]: https://github.github.com/gfm/#atx-headings
1048 [setext headings]: https://github.github.com/gfm/#setext-headings
1049 [inline links]: https://github.github.com/gfm/#inline-link
1050 [reference links]: https://github.github.com/gfm/#reference-link
1051 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1052 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1054 *Matthias St. Pierre*
1056 * The test suite is changed to preserve results of each test recipe.
1057 A new directory test-runs/ with subdirectories named like the
1058 test recipes are created in the build tree for this purpose.
1062 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1063 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1064 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1066 *David von Oheimb, Martin Peylo*
1068 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1069 It supports arbitrary request and response content types, GET redirection,
1070 TLS, connections via HTTP(S) proxies, connections and exchange via
1071 user-defined BIOs (allowing implicit connections), persistent connections,
1072 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1073 The legacy OCSP-focused (and only partly documented) API
1074 is retained for backward compatibility, while most of it is deprecated.
1078 * Added `util/check-format.pl`, a tool for checking adherence to the
1079 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1080 The checks performed are incomplete and yield some false positives.
1081 Still the tool should be useful for detecting most typical glitches.
1085 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1086 If domain name resolution yields multiple IP addresses all of them are tried
1087 after `connect()` failures.
1091 * All of the low level RSA functions have been deprecated.
1095 * X509 certificates signed using SHA1 are no longer allowed at security
1100 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1101 modified to use PKEY APIs. These commands are now in maintenance mode
1102 and no new features will be added to them.
1106 * The command line utility rsautl has been deprecated.
1110 * The command line utilities genrsa and rsa have been modified to use PKEY
1111 APIs. They now write PKCS#8 keys by default. These commands are now in
1112 maintenance mode and no new features will be added to them.
1116 * All of the low level DH functions have been deprecated.
1118 *Paul Dale and Matt Caswell*
1120 * All of the low level DSA functions have been deprecated.
1124 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1125 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1129 * Deprecated low level ECDH and ECDSA functions.
1133 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1137 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1138 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1139 a new formulation to include all the things it can be used for,
1140 as well as words of caution.
1144 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1148 * All of the low level HMAC functions have been deprecated.
1150 *Paul Dale and David von Oheimb*
1152 * Over two thousand fixes were made to the documentation, including:
1153 - Common options (such as -rand/-writerand, TLS version control, etc)
1154 were refactored and point to newly-enhanced descriptions in openssl.pod.
1155 - Added style conformance for all options (with help from Richard Levitte),
1156 documented all reported missing options, added a CI build to check
1157 that all options are documented and that no unimplemented options
1159 - Documented some internals, such as all use of environment variables.
1160 - Addressed all internal broken L<> references.
1164 * All of the low level CMAC functions have been deprecated.
1168 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1169 functions have been deprecated.
1171 *Paul Dale and David von Oheimb*
1173 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1174 set of functions. The documentation mentioned negative values for some
1175 errors, but this was never the case, so the mention of negative values
1178 Code that followed the documentation and thereby check with something
1179 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1183 * All of the low level cipher functions have been deprecated.
1185 *Matt Caswell and Paul Dale*
1187 * Removed include/openssl/opensslconf.h.in and replaced it with
1188 include/openssl/configuration.h.in, which differs in not including
1189 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1190 was added to include both.
1192 This allows internal hacks where one might need to modify the set
1193 of configured macros, for example this if deprecated symbols are
1194 still supposed to be available internally:
1196 #include <openssl/configuration.h>
1198 #undef OPENSSL_NO_DEPRECATED
1199 #define OPENSSL_SUPPRESS_DEPRECATED
1201 #include <openssl/macros.h>
1203 This should not be used by applications that use the exported
1204 symbols, as that will lead to linking errors.
1208 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1209 used in exponentiation with 512-bit moduli. No EC algorithms are
1210 affected. Analysis suggests that attacks against 2-prime RSA1024,
1211 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1212 difficult to perform and are not believed likely. Attacks against DH512
1213 are considered just feasible. However, for an attack the target would
1214 have to re-use the DH512 private key, which is not recommended anyway.
1215 Also applications directly using the low-level API BN_mod_exp may be
1216 affected if they use BN_FLG_CONSTTIME.
1221 * Most memory-debug features have been deprecated, and the functionality
1222 replaced with no-ops.
1226 * Added documentation for the STACK API.
1230 * Introduced a new method type and API, OSSL_ENCODER, to represent
1231 generic encoders. These do the same sort of job that PEM writers
1232 and d2i functions do, but with support for methods supplied by
1233 providers, and the possibility for providers to support other
1238 * Introduced a new method type and API, OSSL_DECODER, to represent
1239 generic decoders. These do the same sort of job that PEM readers
1240 and i2d functions do, but with support for methods supplied by
1241 providers, and the possibility for providers to support other
1246 * Added a .pragma directive to the syntax of configuration files, to
1247 allow varying behavior in a supported and predictable manner.
1248 Currently added pragma:
1252 This allows dollar signs to be a keyword character unless it's
1253 followed by a opening brace or parenthesis. This is useful for
1254 platforms where dollar signs are commonly used in names, such as
1255 volume names and system directory names on VMS.
1259 * Added functionality to create an EVP_PKEY from user data.
1263 * Change the interpretation of the '--api' configuration option to
1264 mean that this is a desired API compatibility level with no
1265 further meaning. The previous interpretation, that this would
1266 also mean to remove all deprecated symbols up to and including
1267 the given version, no requires that 'no-deprecated' is also used
1268 in the configuration.
1270 When building applications, the desired API compatibility level
1271 can be set with the OPENSSL_API_COMPAT macro like before. For
1272 API compatibility version below 3.0, the old style numerical
1273 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1274 For version 3.0 and on, the value is expected to be the decimal
1275 value calculated from the major and minor version like this:
1277 MAJOR * 10000 + MINOR * 100
1281 -DOPENSSL_API_COMPAT=30000 For 3.0
1282 -DOPENSSL_API_COMPAT=30200 For 3.2
1284 To hide declarations that are deprecated up to and including the
1285 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1286 given when building the application as well.
1290 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1291 access to certificate and CRL stores via URIs and OSSL_STORE
1294 This adds the following functions:
1296 - X509_LOOKUP_store()
1297 - X509_STORE_load_file()
1298 - X509_STORE_load_path()
1299 - X509_STORE_load_store()
1300 - SSL_add_store_cert_subjects_to_stack()
1301 - SSL_CTX_set_default_verify_store()
1302 - SSL_CTX_load_verify_file()
1303 - SSL_CTX_load_verify_dir()
1304 - SSL_CTX_load_verify_store()
1308 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1309 The presence of this system service is determined at run-time.
1313 * Added functionality to create an EVP_PKEY context based on data
1314 for methods from providers. This takes an algorithm name and a
1315 property query string and simply stores them, with the intent
1316 that any operation that uses this context will use those strings
1317 to fetch the needed methods implicitly, thereby making the port
1318 of application written for pre-3.0 OpenSSL easier.
1322 * The undocumented function NCONF_WIN32() has been deprecated; for
1323 conversion details see the HISTORY section of doc/man5/config.pod
1327 * Introduced the new functions EVP_DigestSignInit_ex() and
1328 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1329 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1330 pages for further details.
1334 * Over two thousand fixes were made to the documentation, including:
1335 adding missing command flags, better style conformance, documentation
1338 *Rich Salz, Richard Levitte*
1340 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1341 X25519, X448, Ed25519 and Ed448.
1345 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1350 * Deprecated the public definition of `ERR_STATE` as well as the function
1351 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1356 * Added ERR functionality to give callers access to the stored function
1357 names that have replaced the older function code based functions.
1359 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1360 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1361 ERR_peek_error_all() and ERR_peek_last_error_all().
1363 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1364 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1365 ERR_func_error_string().
1369 * Extended testing to be verbose for failing tests only. The make variables
1370 VERBOSE_FAILURE or VF can be used to enable this:
1372 $ make VF=1 test # Unix
1373 $ mms /macro=(VF=1) test ! OpenVMS
1374 $ nmake VF=1 test # Windows
1378 * Added the `-copy_extensions` option to the `x509` command for use with
1379 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1380 all extensions in the request are copied to the certificate or vice versa.
1382 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1384 * Added the `-copy_extensions` option to the `req` command for use with
1385 `-x509`. When given with the `copy` or `copyall` argument,
1386 all extensions in the certification request are copied to the certificate.
1390 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1391 they generate are by default RFC 5280 compliant in the following sense:
1392 There is a subjectKeyIdentifier extension with a hash value of the public key
1393 and for not self-signed certs there is an authorityKeyIdentifier extension
1394 with a keyIdentifier field or issuer information identifying the signing key.
1395 This is done unless some configuration overrides the new default behavior,
1396 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1400 * Added several checks to `X509_verify_cert()` according to requirements in
1401 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1402 (which may be done by using the CLI option `-x509_strict`):
1403 * The basicConstraints of CA certificates must be marked critical.
1404 * CA certificates must explicitly include the keyUsage extension.
1405 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1406 * The issuer name of any certificate must not be empty.
1407 * The subject name of CA certs, certs with keyUsage crlSign,
1408 and certs without subjectAlternativeName must not be empty.
1409 * If a subjectAlternativeName extension is given it must not be empty.
1410 * The signatureAlgorithm field and the cert signature must be consistent.
1411 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1412 must not be marked critical.
1413 * The authorityKeyIdentifier must be given for X.509v3 certs
1414 unless they are self-signed.
1415 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1419 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1420 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1424 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1425 used even when parsing explicit parameters, when loading a encoded key
1426 or calling `EC_GROUP_new_from_ecpkparameters()`/
1427 `EC_GROUP_new_from_ecparameters()`.
1428 This prevents bypass of security hardening and performance gains,
1429 especially for curves with specialized EC_METHODs.
1430 By default, if a key encoded with explicit parameters is loaded and later
1431 encoded, the output is still encoded with explicit parameters, even if
1432 internally a "named" EC_GROUP is used for computation.
1436 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1437 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1438 NULL. After this change, only the cofactor parameter can be NULL. It also
1439 does some minimal sanity checks on the passed order.
1444 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1445 An attack is simple, if the first CMS_recipientInfo is valid but the
1446 second CMS_recipientInfo is chosen ciphertext. If the second
1447 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1448 encryption key will be replaced by garbage, and the message cannot be
1449 decoded, but if the RSA decryption fails, the correct encryption key is
1450 used and the recipient will not notice the attack.
1451 As a work around for this potential attack the length of the decrypted
1452 key must be equal to the cipher default key length, in case the
1453 certifiate is not given and all recipientInfo are tried out.
1454 The old behaviour can be re-enabled in the CMS code by setting the
1455 CMS_DEBUG_DECRYPT flag.
1459 * Early start up entropy quality from the DEVRANDOM seed source has been
1460 improved for older Linux systems. The RAND subsystem will wait for
1461 /dev/random to be producing output before seeding from /dev/urandom.
1462 The seeded state is stored for future library initialisations using
1463 a system global shared memory segment. The shared memory identifier
1464 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1465 the desired value. The default identifier is 114.
1469 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1470 when primes for RSA keys are computed.
1471 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1472 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1473 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1474 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1475 This avoids possible fingerprinting of newly generated RSA modules.
1479 * Correct the extended master secret constant on EBCDIC systems. Without this
1480 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1481 negotiate EMS will fail. Unfortunately this also means that TLS connections
1482 between EBCDIC systems with this fix, and EBCDIC systems without this
1483 fix will fail if they negotiate EMS.
1487 * Changed the library initialisation so that the config file is now loaded
1488 by default. This was already the case for libssl. It now occurs for both
1489 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1490 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1494 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1495 where the former acts as a replacement for `ERR_put_error()`, and the
1496 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1497 `ERR_raise_data()` adds more flexibility by taking a format string and
1498 an arbitrary number of arguments following it, to be processed with
1503 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1504 to check if a named provider is loaded and available. When called, it
1505 will also activate all fallback providers if such are still present.
1509 * Enforce a minimum DH modulus size of 512 bits.
1513 * Changed DH parameters to generate the order q subgroup instead of 2q.
1514 Previously generated DH parameters are still accepted by DH_check
1515 but DH_generate_key works around that by clearing bit 0 of the
1516 private key for those. This avoids leaking bit 0 of the private key.
1520 * Significantly reduce secure memory usage by the randomness pools.
1524 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1529 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1530 algorithms. An implementation of a key exchange algorithm can be obtained
1531 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1532 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1533 the older EVP_PKEY_derive_init() function. See the man pages for the new
1534 functions for further details.
1538 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1542 * Removed the function names from error messages and deprecated the
1547 * Removed NextStep support and the macro OPENSSL_UNISTD
1551 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1552 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1553 Also removed "export var as function" capability; we do not export
1554 variables, only functions.
1558 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1559 an error and 1 indicating success. In previous versions of OpenSSL this
1560 was a void type. If a key was set longer than the maximum possible this
1565 * Support SM2 signing and verification schemes with X509 certificate.
1569 * Use SHA256 as the default digest for TS query in the `ts` app.
1573 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1577 * Default cipher lists/suites are now available via a function, the
1578 #defines are deprecated.
1582 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1583 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1584 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1588 * Join the directories crypto/x509 and crypto/x509v3
1592 * Added command 'openssl kdf' that uses the EVP_KDF API.
1596 * Added command 'openssl mac' that uses the EVP_MAC API.
1600 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1601 as default directories. Also added the command 'openssl info'
1602 for scripting purposes.
1606 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1611 * Add prediction resistance to the DRBG reseeding process.
1615 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1616 mandated by IEEE Std 1619-2018.
1620 * Added newline escaping functionality to a filename when using openssl dgst.
1621 This output format is to replicate the output format found in the `*sum`
1622 checksum programs. This aims to preserve backward compatibility.
1624 *Matt Eaton, Richard Levitte, and Paul Dale*
1626 * Removed the heartbeat message in DTLS feature, as it has very
1627 little usage and doesn't seem to fulfill a valuable purpose.
1628 The configuration option is now deprecated.
1632 * Changed the output of 'openssl {digestname} < file' to display the
1633 digest name in its output.
1637 * Added a new generic trace API which provides support for enabling
1638 instrumentation through trace output.
1640 *Richard Levitte & Matthias St. Pierre*
1642 * Added build tests for C++. These are generated files that only do one
1643 thing, to include one public OpenSSL head file each. This tests that
1644 the public header files can be usefully included in a C++ application.
1646 This test isn't enabled by default. It can be enabled with the option
1647 'enable-buildtest-c++'.
1651 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1655 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1659 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1663 * Added KMAC to EVP_MAC.
1667 * Added property based algorithm implementation selection framework to
1672 * Added SCA hardening for modular field inversion in EC_GROUP through
1673 a new dedicated field_inv() pointer in EC_METHOD.
1674 This also addresses a leakage affecting conversions from projective
1675 to affine coordinates.
1677 *Billy Bob Brumley, Nicola Tuveri*
1679 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1680 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1681 those algorithms that were already supported through the EVP_PKEY API
1682 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1683 and scrypt are now wrappers that call EVP_KDF.
1687 * Build devcrypto engine as a dynamic engine.
1689 *Eneas U de Queiroz*
1691 * Add keyed BLAKE2 to EVP_MAC.
1695 * Fix a bug in the computation of the endpoint-pair shared secret used
1696 by DTLS over SCTP. This breaks interoperability with older versions
1697 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1698 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1699 interoperability with such broken implementations. However, enabling
1700 this switch breaks interoperability with correct implementations.
1702 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1703 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1707 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1711 * Changed the license to the Apache License v2.0.
1715 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1717 - Major releases (indicated by incrementing the MAJOR release number)
1718 may introduce incompatible API/ABI changes.
1719 - Minor releases (indicated by incrementing the MINOR release number)
1720 may introduce new features but retain API/ABI compatibility.
1721 - Patch releases (indicated by incrementing the PATCH number)
1722 are intended for bug fixes and other improvements of existing
1723 features only (like improving performance or adding documentation)
1724 and retain API/ABI compatibility.
1728 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1732 * Remove the 'dist' target and add a tarball building script. The
1733 'dist' target has fallen out of use, and it shouldn't be
1734 necessary to configure just to create a source distribution.
1738 * Recreate the OS390-Unix config target. It no longer relies on a
1739 special script like it did for OpenSSL pre-1.1.0.
1743 * Instead of having the source directories listed in Configure, add
1744 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1749 * Add GMAC to EVP_MAC.
1753 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1757 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1758 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1759 to facilitate the continued use of MACs through raw private keys in
1760 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1764 * Deprecate ECDH_KDF_X9_62().
1768 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1769 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1770 are retained for backwards compatibility.
1774 * AES-XTS mode now enforces that its two keys are different to mitigate
1775 the attacked described in "Efficient Instantiations of Tweakable
1776 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1777 Details of this attack can be obtained from:
1778 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1782 * Rename the object files, i.e. give them other names than in previous
1783 versions. Their names now include the name of the final product, as
1784 well as its type mnemonic (bin, lib, shlib).
1788 * Added new option for 'openssl list', '-objects', which will display the
1789 list of built in objects, i.e. OIDs with names.
1793 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1794 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1799 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1800 improves application performance by removing data copies and providing
1801 applications with zero-copy system calls such as sendfile and splice.
1805 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1809 * `PKCS12_parse` now maintains the order of the parsed certificates
1810 when outputting them via `*ca` (rather than reversing it).
1814 * Deprecated pthread fork support methods.
1818 * Added support for FFDHE key exchange in TLS 1.3.
1822 * Added a new concept for OpenSSL plugability: providers. This
1823 functionality is designed to replace the ENGINE API and ENGINE
1824 implementations, and to be much more dynamic, allowing provider
1825 authors to introduce new algorithms among other things, as long as
1826 there's an API that supports the algorithm type.
1828 With this concept comes a new core API for interaction between
1829 libcrypto and provider implementations. Public libcrypto functions
1830 that want to use providers do so through this core API.
1832 The main documentation for this core API is found in
1833 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1834 refer to other manuals describing the API specific for supported
1835 algorithm types (also called operations).
1842 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
1844 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
1846 * Avoid loading of a dynamic engine twice.
1850 * Prioritise DANE TLSA issuer certs over peer certs
1854 * Fixed random API for MacOS prior to 10.12
1856 These MacOS versions don't support the CommonCrypto APIs
1860 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1862 * Fixed an SM2 Decryption Buffer Overflow.
1864 In order to decrypt SM2 encrypted data an application is expected to
1865 call the API function EVP_PKEY_decrypt(). Typically an application will
1866 call this function twice. The first time, on entry, the "out" parameter
1867 can be NULL and, on exit, the "outlen" parameter is populated with the
1868 buffer size required to hold the decrypted plaintext. The application
1869 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1870 again, but this time passing a non-NULL value for the "out" parameter.
1872 A bug in the implementation of the SM2 decryption code means that the
1873 calculation of the buffer size required to hold the plaintext returned
1874 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1875 size required by the second call. This can lead to a buffer overflow
1876 when EVP_PKEY_decrypt() is called by the application a second time with
1877 a buffer that is too small.
1879 A malicious attacker who is able present SM2 content for decryption to
1880 an application could cause attacker chosen data to overflow the buffer
1881 by up to a maximum of 62 bytes altering the contents of other data held
1882 after the buffer, possibly changing application behaviour or causing
1883 the application to crash. The location of the buffer is application
1884 dependent but is typically heap allocated.
1889 * Fixed various read buffer overruns processing ASN.1 strings
1891 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1892 structure which contains a buffer holding the string data and a field
1893 holding the buffer length. This contrasts with normal C strings which
1894 are repesented as a buffer for the string data which is terminated
1895 with a NUL (0) byte.
1897 Although not a strict requirement, ASN.1 strings that are parsed using
1898 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1899 well as any string whose value has been set with the ASN1_STRING_set()
1900 function will additionally NUL terminate the byte array in the
1901 ASN1_STRING structure.
1903 However, it is possible for applications to directly construct valid
1904 ASN1_STRING structures which do not NUL terminate the byte array by
1905 directly setting the "data" and "length" fields in the ASN1_STRING
1906 array. This can also happen by using the ASN1_STRING_set0() function.
1908 Numerous OpenSSL functions that print ASN.1 data have been found to
1909 assume that the ASN1_STRING byte array will be NUL terminated, even
1910 though this is not guaranteed for strings that have been directly
1911 constructed. Where an application requests an ASN.1 structure to be
1912 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1913 been directly constructed by the application without NUL terminating
1914 the "data" field, then a read buffer overrun can occur.
1916 The same thing can also occur during name constraints processing
1917 of certificates (for example if a certificate has been directly
1918 constructed by the application instead of loading it via the OpenSSL
1919 parsing functions, and the certificate contains non NUL terminated
1920 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1921 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1923 If a malicious actor can cause an application to directly construct an
1924 ASN1_STRING and then process it through one of the affected OpenSSL
1925 functions then this issue could be hit. This might result in a crash
1926 (causing a Denial of Service attack). It could also result in the
1927 disclosure of private memory contents (such as private keys, or
1928 sensitive plaintext).
1933 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1935 * Fixed a problem with verifying a certificate chain when using the
1936 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1937 the certificates present in a certificate chain. It is not set by default.
1939 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1940 the chain that have explicitly encoded elliptic curve parameters was added
1941 as an additional strict check.
1943 An error in the implementation of this check meant that the result of a
1944 previous check to confirm that certificates in the chain are valid CA
1945 certificates was overwritten. This effectively bypasses the check
1946 that non-CA certificates must not be able to issue other certificates.
1948 If a "purpose" has been configured then there is a subsequent opportunity
1949 for checks that the certificate is a valid CA. All of the named "purpose"
1950 values implemented in libcrypto perform this check. Therefore, where
1951 a purpose is set the certificate chain will still be rejected even when the
1952 strict flag has been used. A purpose is set by default in libssl client and
1953 server certificate verification routines, but it can be overridden or
1954 removed by an application.
1956 In order to be affected, an application must explicitly set the
1957 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1958 for the certificate verification or, in the case of TLS client or server
1959 applications, override the default purpose.
1964 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1965 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1966 renegotiation ClientHello omits the signature_algorithms extension (where it
1967 was present in the initial ClientHello), but includes a
1968 signature_algorithms_cert extension then a NULL pointer dereference will
1969 result, leading to a crash and a denial of service attack.
1971 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1972 (which is the default configuration). OpenSSL TLS clients are not impacted by
1976 *Peter Kästle and Samuel Sapalski*
1978 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1980 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1981 create a unique hash value based on the issuer and serial number data
1982 contained within an X509 certificate. However it was failing to correctly
1983 handle any errors that may occur while parsing the issuer field (which might
1984 occur if the issuer field is maliciously constructed). This may subsequently
1985 result in a NULL pointer deref and a crash leading to a potential denial of
1991 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1992 padding mode to correctly check for rollback attacks. This is considered a
1993 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1998 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1999 functions. Previously they could overflow the output length argument in some
2000 cases where the input length is close to the maximum permissable length for
2001 an integer on the platform. In such cases the return value from the function
2002 call would be 1 (indicating success), but the output length value would be
2003 negative. This could cause applications to behave incorrectly or crash.
2008 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2009 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2010 could be exploited in a side channel attack to recover the password. Since
2011 the attack is local host only this is outside of the current OpenSSL
2012 threat model and therefore no CVE is assigned.
2014 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2019 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2021 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2022 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2023 If an attacker can control both items being compared then this could lead
2024 to a possible denial of service attack. OpenSSL itself uses the
2025 GENERAL_NAME_cmp function for two purposes:
2026 1) Comparing CRL distribution point names between an available CRL and a
2027 CRL distribution point embedded in an X509 certificate
2028 2) When verifying that a timestamp response token signer matches the
2029 timestamp authority name (exposed via the API functions
2030 TS_RESP_verify_response and TS_RESP_verify_token)
2035 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2037 * Certificates with explicit curve parameters are now disallowed in
2038 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2042 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2043 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2044 conversely, silently ignore DTLS protocol version bounds when configuring
2045 TLS-based contexts. The commands can be repeated to set bounds of both
2046 types. The same applies with the corresponding "min_protocol" and
2047 "max_protocol" command-line switches, in case some application uses both TLS
2050 SSL_CTX instances that are created for a fixed protocol version (e.g.
2051 TLSv1_server_method()) also silently ignore version bounds. Previously
2052 attempts to apply bounds to these protocol versions would result in an
2053 error. Now only the "version-flexible" SSL_CTX instances are subject to
2054 limits in configuration files in command-line options.
2058 * Handshake now fails if Extended Master Secret extension is dropped
2063 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2065 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2067 * Fixed segmentation fault in SSL_check_chain()
2068 Server or client applications that call the SSL_check_chain() function
2069 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2070 dereference as a result of incorrect handling of the
2071 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2072 or unrecognised signature algorithm is received from the peer. This could
2073 be exploited by a malicious peer in a Denial of Service attack.
2078 * Added AES consttime code for no-asm configurations
2079 an optional constant time support for AES was added
2080 when building openssl for no-asm.
2081 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2082 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2083 At this time this feature is by default disabled.
2084 It will be enabled by default in 3.0.
2088 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2090 * Revert the change of EOF detection while reading in libssl to avoid
2091 regressions in applications depending on the current way of reporting
2092 the EOF. As the existing method is not fully accurate the change to
2093 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2094 branch and will be present in the 3.0 release.
2098 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2099 when primes for RSA keys are computed.
2100 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2101 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2102 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2103 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2104 This avoids possible fingerprinting of newly generated RSA modules.
2108 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2110 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2111 while reading in libssl then we would report an error back to the
2112 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2113 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2114 therefore give a hint as to what went wrong.
2118 * Check that ed25519 and ed448 are allowed by the security level. Previously
2119 signature algorithms not using an MD were not being checked that they were
2120 allowed by the security level.
2124 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2125 was not quite right. The behaviour was not consistent between resumption
2126 and normal handshakes, and also not quite consistent with historical
2127 behaviour. The behaviour in various scenarios has been clarified and
2128 it has been updated to make it match historical behaviour as closely as
2133 * *[VMS only]* The header files that the VMS compilers include automatically,
2134 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2135 that the C++ compiler doesn't understand. This is a shortcoming in the
2136 compiler, but can be worked around with `__cplusplus` guards.
2138 C++ applications that use OpenSSL libraries must be compiled using the
2139 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2140 functions. Otherwise, only functions with symbols of less than 31
2141 characters can be used, as the linker will not be able to successfully
2142 resolve symbols with longer names.
2146 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2147 The presence of this system service is determined at run-time.
2151 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2156 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2158 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2159 number generator (RNG). This was intended to include protection in the
2160 event of a fork() system call in order to ensure that the parent and child
2161 processes did not share the same RNG state. However this protection was not
2162 being used in the default case.
2164 A partial mitigation for this issue is that the output from a high
2165 precision timer is mixed into the RNG state so the likelihood of a parent
2166 and child process sharing state is significantly reduced.
2168 If an application already calls OPENSSL_init_crypto() explicitly using
2169 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2172 *Matthias St. Pierre*
2174 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2175 used even when parsing explicit parameters, when loading a encoded key
2176 or calling `EC_GROUP_new_from_ecpkparameters()`/
2177 `EC_GROUP_new_from_ecparameters()`.
2178 This prevents bypass of security hardening and performance gains,
2179 especially for curves with specialized EC_METHODs.
2180 By default, if a key encoded with explicit parameters is loaded and later
2181 encoded, the output is still encoded with explicit parameters, even if
2182 internally a "named" EC_GROUP is used for computation.
2186 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2187 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2188 NULL. After this change, only the cofactor parameter can be NULL. It also
2189 does some minimal sanity checks on the passed order.
2194 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2195 An attack is simple, if the first CMS_recipientInfo is valid but the
2196 second CMS_recipientInfo is chosen ciphertext. If the second
2197 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2198 encryption key will be replaced by garbage, and the message cannot be
2199 decoded, but if the RSA decryption fails, the correct encryption key is
2200 used and the recipient will not notice the attack.
2201 As a work around for this potential attack the length of the decrypted
2202 key must be equal to the cipher default key length, in case the
2203 certifiate is not given and all recipientInfo are tried out.
2204 The old behaviour can be re-enabled in the CMS code by setting the
2205 CMS_DEBUG_DECRYPT flag.
2210 * Early start up entropy quality from the DEVRANDOM seed source has been
2211 improved for older Linux systems. The RAND subsystem will wait for
2212 /dev/random to be producing output before seeding from /dev/urandom.
2213 The seeded state is stored for future library initialisations using
2214 a system global shared memory segment. The shared memory identifier
2215 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2216 the desired value. The default identifier is 114.
2220 * Correct the extended master secret constant on EBCDIC systems. Without this
2221 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2222 negotiate EMS will fail. Unfortunately this also means that TLS connections
2223 between EBCDIC systems with this fix, and EBCDIC systems without this
2224 fix will fail if they negotiate EMS.
2228 * Use Windows installation paths in the mingw builds
2230 Mingw isn't a POSIX environment per se, which means that Windows
2231 paths should be used for installation.
2236 * Changed DH_check to accept parameters with order q and 2q subgroups.
2237 With order 2q subgroups the bit 0 of the private key is not secret
2238 but DH_generate_key works around that by clearing bit 0 of the
2239 private key for those. This avoids leaking bit 0 of the private key.
2243 * Significantly reduce secure memory usage by the randomness pools.
2247 * Revert the DEVRANDOM_WAIT feature for Linux systems
2249 The DEVRANDOM_WAIT feature added a select() call to wait for the
2250 /dev/random device to become readable before reading from the
2251 /dev/urandom device.
2253 It turned out that this change had negative side effects on
2254 performance which were not acceptable. After some discussion it
2255 was decided to revert this feature and leave it up to the OS
2256 resp. the platform maintainer to ensure a proper initialization
2257 during early boot time.
2259 *Matthias St. Pierre*
2261 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2263 * Add build tests for C++. These are generated files that only do one
2264 thing, to include one public OpenSSL head file each. This tests that
2265 the public header files can be usefully included in a C++ application.
2267 This test isn't enabled by default. It can be enabled with the option
2268 'enable-buildtest-c++'.
2272 * Enable SHA3 pre-hashing for ECDSA and DSA.
2276 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2277 This changes the size when using the `genpkey` command when no size is given.
2278 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2279 generation commands to use 2048 bits by default.
2283 * Reorganize the manual pages to consistently have RETURN VALUES,
2284 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2285 util/fix-doc-nits accordingly.
2287 *Paul Yang, Joshua Lock*
2289 * Add the missing accessor EVP_PKEY_get0_engine()
2293 * Have commands like `s_client` and `s_server` output the signature scheme
2294 along with other cipher suite parameters when debugging.
2298 * Make OPENSSL_config() error agnostic again.
2302 * Do the error handling in RSA decryption constant time.
2306 * Prevent over long nonces in ChaCha20-Poly1305.
2308 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2309 for every encryption operation. RFC 7539 specifies that the nonce value
2310 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2311 and front pads the nonce with 0 bytes if it is less than 12
2312 bytes. However it also incorrectly allows a nonce to be set of up to 16
2313 bytes. In this case only the last 12 bytes are significant and any
2314 additional leading bytes are ignored.
2316 It is a requirement of using this cipher that nonce values are
2317 unique. Messages encrypted using a reused nonce value are susceptible to
2318 serious confidentiality and integrity attacks. If an application changes
2319 the default nonce length to be longer than 12 bytes and then makes a
2320 change to the leading bytes of the nonce expecting the new value to be a
2321 new unique nonce then such an application could inadvertently encrypt
2322 messages with a reused nonce.
2324 Additionally the ignored bytes in a long nonce are not covered by the
2325 integrity guarantee of this cipher. Any application that relies on the
2326 integrity of these ignored leading bytes of a long nonce may be further
2327 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2328 is safe because no such use sets such a long nonce value. However user
2329 applications that use this cipher directly and set a non-default nonce
2330 length to be longer than 12 bytes may be vulnerable.
2332 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2338 * Add DEVRANDOM_WAIT feature for Linux systems
2340 On older Linux systems where the getrandom() system call is not available,
2341 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2342 Contrary to getrandom(), the /dev/urandom device will not block during
2343 early boot when the kernel CSPRNG has not been seeded yet.
2345 To mitigate this known weakness, use select() to wait for /dev/random to
2346 become readable before reading from /dev/urandom.
2348 * Ensure that SM2 only uses SM3 as digest algorithm
2352 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2354 * Change the info callback signals for the start and end of a post-handshake
2355 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2356 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2357 confused by this and assume that a TLSv1.2 renegotiation has started. This
2358 can break KeyUpdate handling. Instead we no longer signal the start and end
2359 of a post handshake message exchange (although the messages themselves are
2360 still signalled). This could break some applications that were expecting
2361 the old signals. However without this KeyUpdate is not usable for many
2366 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2368 * Timing vulnerability in DSA signature generation
2370 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2371 timing side channel attack. An attacker could use variations in the signing
2372 algorithm to recover the private key.
2374 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2379 * Timing vulnerability in ECDSA signature generation
2381 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2382 timing side channel attack. An attacker could use variations in the signing
2383 algorithm to recover the private key.
2385 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2390 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2391 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2392 of two gigabytes and the error handling improved.
2394 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2395 categorized as a normal bug, not a security issue, because the DRBG reseeds
2396 automatically and is fully functional even without additional randomness
2397 provided by the application.
2399 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2401 * Add a new ClientHello callback. Provides a callback interface that gives
2402 the application the ability to adjust the nascent SSL object at the
2403 earliest stage of ClientHello processing, immediately after extensions have
2404 been collected but before they have been processed. In particular, this
2405 callback can adjust the supported TLS versions in response to the contents
2410 * Add SM2 base algorithm support.
2414 * s390x assembly pack: add (improved) hardware-support for the following
2415 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2416 aes-cfb/cfb8, aes-ecb.
2420 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2421 parameter is no longer accepted, as it leads to a corrupt table. NULL
2422 pem_str is reserved for alias entries only.
2426 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2427 step for prime curves. The new implementation is based on formulae from
2428 differential addition-and-doubling in homogeneous projective coordinates
2429 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2430 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2431 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2432 to work in projective coordinates.
2434 *Billy Bob Brumley, Nicola Tuveri*
2436 * Change generating and checking of primes so that the error rate of not
2437 being prime depends on the intended use based on the size of the input.
2438 For larger primes this will result in more rounds of Miller-Rabin.
2439 The maximal error rate for primes with more than 1080 bits is lowered
2442 *Kurt Roeckx, Annie Yousar*
2444 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2448 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2449 moving between systems, and to avoid confusion when a Windows build is
2450 done with mingw vs with MSVC. For POSIX installs, there's still a
2451 symlink or copy named 'tsget' to avoid that confusion as well.
2455 * Revert blinding in ECDSA sign and instead make problematic addition
2456 length-invariant. Switch even to fixed-length Montgomery multiplication.
2460 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2461 step for binary curves. The new implementation is based on formulae from
2462 differential addition-and-doubling in mixed Lopez-Dahab projective
2463 coordinates, modified to independently blind the operands.
2465 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2467 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2468 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2469 EC_METHODs to implement their own specialized "ladder step", to take
2470 advantage of more favorable coordinate systems or more efficient
2471 differential addition-and-doubling algorithms.
2473 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2475 * Modified the random device based seed sources to keep the relevant
2476 file descriptors open rather than reopening them on each access.
2477 This allows such sources to operate in a chroot() jail without
2478 the associated device nodes being available. This behaviour can be
2479 controlled using RAND_keep_random_devices_open().
2483 * Numerous side-channel attack mitigations have been applied. This may have
2484 performance impacts for some algorithms for the benefit of improved
2485 security. Specific changes are noted in this change log by their respective
2490 * AIX shared library support overhaul. Switch to AIX "natural" way of
2491 handling shared libraries, which means collecting shared objects of
2492 different versions and bitnesses in one common archive. This allows to
2493 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2494 doesn't affect the way 3rd party applications are linked, only how
2495 multi-version installation is managed.
2499 * Make ec_group_do_inverse_ord() more robust and available to other
2500 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2501 mitigations are applied to the fallback BN_mod_inverse().
2502 When using this function rather than BN_mod_inverse() directly, new
2503 EC cryptosystem implementations are then safer-by-default.
2507 * Add coordinate blinding for EC_POINT and implement projective
2508 coordinate blinding for generic prime curves as a countermeasure to
2509 chosen point SCA attacks.
2511 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2513 * Add blinding to ECDSA and DSA signatures to protect against side channel
2514 attacks discovered by Keegan Ryan (NCC Group).
2518 * Enforce checking in the `pkeyutl` command to ensure that the input
2519 length does not exceed the maximum supported digest length when performing
2520 a sign, verify or verifyrecover operation.
2524 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2525 I/O in combination with something like select() or poll() will hang. This
2526 can be turned off again using SSL_CTX_clear_mode().
2527 Many applications do not properly handle non-application data records, and
2528 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2529 around the problems in those applications, but can also break some.
2530 It's recommended to read the manpages about SSL_read(), SSL_write(),
2531 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2532 SSL_CTX_set_read_ahead() again.
2536 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2537 now allow empty (zero character) pass phrases.
2541 * Apply blinding to binary field modular inversion and remove patent
2542 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2546 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2547 binary and prime elliptic curves.
2551 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2552 constant time fixed point multiplication.
2556 * Revise elliptic curve scalar multiplication with timing attack
2557 defenses: ec_wNAF_mul redirects to a constant time implementation
2558 when computing fixed point and variable point multiplication (which
2559 in OpenSSL are mostly used with secret scalars in keygen, sign,
2560 ECDH derive operations).
2561 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2564 * Updated CONTRIBUTING
2568 * Updated DRBG / RAND to request nonce and additional low entropy
2569 randomness from the system.
2571 *Matthias St. Pierre*
2573 * Updated 'openssl rehash' to use OpenSSL consistent default.
2577 * Moved the load of the ssl_conf module to libcrypto, which helps
2578 loading engines that libssl uses before libssl is initialised.
2582 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2586 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2588 *Ingo Schwarze, Rich Salz*
2590 * Added output of accepting IP address and port for 'openssl s_server'
2594 * Added a new API for TLSv1.3 ciphersuites:
2595 SSL_CTX_set_ciphersuites()
2596 SSL_set_ciphersuites()
2600 * Memory allocation failures consistently add an error to the error
2605 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2606 in libcrypto when run as setuid/setgid.
2610 * Load any config file by default when libssl is used.
2614 * Added new public header file <openssl/rand_drbg.h> and documentation
2615 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2617 *Matthias St. Pierre*
2619 * QNX support removed (cannot find contributors to get their approval
2620 for the license change).
2624 * TLSv1.3 replay protection for early data has been implemented. See the
2625 SSL_read_early_data() man page for further details.
2629 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2630 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2631 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2632 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2633 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2634 configuration has been separated out. See the ciphers man page or the
2635 SSL_CTX_set_ciphersuites() man page for more information.
2639 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2640 in responder mode now supports the new "-multi" option, which
2641 spawns the specified number of child processes to handle OCSP
2642 requests. The "-timeout" option now also limits the OCSP
2643 responder's patience to wait to receive the full client request
2644 on a newly accepted connection. Child processes are respawned
2645 as needed, and the CA index file is automatically reloaded
2646 when changed. This makes it possible to run the "ocsp" responder
2647 as a long-running service, making the OpenSSL CA somewhat more
2648 feature-complete. In this mode, most diagnostic messages logged
2649 after entering the event loop are logged via syslog(3) rather than
2654 * Added support for X448 and Ed448. Heavily based on original work by
2659 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2660 objects loaded. This adds the functions OSSL_STORE_expect() and
2661 OSSL_STORE_find() as well as needed tools to construct searches and
2662 get the search data out of them.
2666 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2667 version of OpenSSL should review their configuration settings to ensure
2668 that they are still appropriate for TLSv1.3. For further information see:
2669 <https://wiki.openssl.org/index.php/TLS1.3>
2673 * Grand redesign of the OpenSSL random generator
2675 The default RAND method now utilizes an AES-CTR DRBG according to
2676 NIST standard SP 800-90Ar1. The new random generator is essentially
2677 a port of the default random generator from the OpenSSL FIPS 2.0
2678 object module. It is a hybrid deterministic random bit generator
2679 using an AES-CTR bit stream and which seeds and reseeds itself
2680 automatically using trusted system entropy sources.
2682 Some of its new features are:
2683 - Support for multiple DRBG instances with seed chaining.
2684 - The default RAND method makes use of a DRBG.
2685 - There is a public and private DRBG instance.
2686 - The DRBG instances are fork-safe.
2687 - Keep all global DRBG instances on the secure heap if it is enabled.
2688 - The public and private DRBG instance are per thread for lock free
2691 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2693 * Changed Configure so it only says what it does and doesn't dump
2694 so much data. Instead, ./configdata.pm should be used as a script
2695 to display all sorts of configuration data.
2699 * Added processing of "make variables" to Configure.
2703 * Added SHA512/224 and SHA512/256 algorithm support.
2707 * The last traces of Netware support, first removed in 1.1.0, have
2712 * Get rid of Makefile.shared, and in the process, make the processing
2713 of certain files (rc.obj, or the .def/.map/.opt files produced from
2714 the ordinal files) more visible and hopefully easier to trace and
2715 debug (or make silent).
2719 * Make it possible to have environment variable assignments as
2720 arguments to config / Configure.
2724 * Add multi-prime RSA (RFC 8017) support.
2728 * Add SM3 implemented according to GB/T 32905-2016
2729 *Jack Lloyd <jack.lloyd@ribose.com>,*
2730 *Ronald Tse <ronald.tse@ribose.com>,*
2731 *Erick Borsboom <erick.borsboom@ribose.com>*
2733 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2734 as documented in RFC6066.
2735 Based on a patch from Tomasz Moń
2737 *Filipe Raimundo da Silva*
2739 * Add SM4 implemented according to GB/T 32907-2016.
2740 *Jack Lloyd <jack.lloyd@ribose.com>,*
2741 *Ronald Tse <ronald.tse@ribose.com>,*
2742 *Erick Borsboom <erick.borsboom@ribose.com>*
2744 * Reimplement -newreq-nodes and ERR_error_string_n; the
2745 original author does not agree with the license change.
2749 * Add ARIA AEAD TLS support.
2753 * Some macro definitions to support VS6 have been removed. Visual
2754 Studio 6 has not worked since 1.1.0
2758 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2759 without clearing the errors.
2763 * Add "atfork" functions. If building on a system that without
2764 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2765 requirements. The RAND facility now uses/requires this.
2773 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2774 not possible to disable entirely. However, it's still possible to
2775 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2778 To disable, configure with 'no-ui-console'. 'no-ui' is still
2779 possible to use as an alias. Check at compile time with the
2780 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2781 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2785 * Add a STORE module, which implements a uniform and URI based reader of
2786 stores that can contain keys, certificates, CRLs and numerous other
2787 objects. The main API is loosely based on a few stdio functions,
2788 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2789 OSSL_STORE_error and OSSL_STORE_close.
2790 The implementation uses backends called "loaders" to implement arbitrary
2791 URI schemes. There is one built in "loader" for the 'file' scheme.
2795 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2796 then adjusted to work on FreeBSD 8.4 as well.
2797 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2798 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2802 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2803 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2804 error code calls like this:
2806 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2808 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2809 that can be encoded in C. For the foreseeable future, this will only
2812 *Richard Levitte and Tim Hudson*
2814 * Removed BSD cryptodev engine.
2818 * Add a build target 'build_all_generated', to build all generated files
2819 and only that. This can be used to prepare everything that requires
2820 things like perl for a system that lacks perl and then move everything
2821 to that system and do the rest of the build there.
2825 * In the UI interface, make it possible to duplicate the user data. This
2826 can be used by engines that need to retain the data for a longer time
2827 than just the call where this user data is passed.
2831 * Ignore the '-named_curve auto' value for compatibility of applications
2834 *Tomáš Mráz <tmraz@fedoraproject.org>*
2836 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2837 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2838 alerts across multiple records (some of which could be empty). In practice
2839 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2840 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2841 support this at all. Supporting it adds significant complexity to the
2842 record layer, and its removal is unlikely to cause interoperability
2847 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2848 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2849 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2854 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2855 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2857 *Richard Levitte, Andy Polyakov*
2859 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2864 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2865 platform rather than 'mingw'.
2869 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2870 success if they are asked to add an object which already exists
2871 in the store. This change cascades to other functions which load
2872 certificates and CRLs.
2876 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2877 facilitate stack unwinding even from assembly subroutines.
2881 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2882 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2886 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2887 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2888 which is the minimum version we support.
2892 * Certificate time validation (X509_cmp_time) enforces stricter
2893 compliance with RFC 5280. Fractional seconds and timezone offsets
2894 are no longer allowed.
2898 * Add support for ARIA
2902 * s_client will now send the Server Name Indication (SNI) extension by
2903 default unless the new "-noservername" option is used. The server name is
2904 based on the host provided to the "-connect" option unless overridden by
2905 using "-servername".
2909 * Add support for SipHash
2913 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2914 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2915 prevent issues where no progress is being made and the peer continually
2916 sends unrecognised record types, using up resources processing them.
2920 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2921 using the algorithm defined in
2922 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2926 * Heartbeat support has been removed; the ABI is changed for now.
2928 *Richard Levitte, Rich Salz*
2930 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2934 * The RSA "null" method, which was partially supported to avoid patent
2935 issues, has been replaced to always returns NULL.
2942 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2944 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2945 used even when parsing explicit parameters, when loading a encoded key
2946 or calling `EC_GROUP_new_from_ecpkparameters()`/
2947 `EC_GROUP_new_from_ecparameters()`.
2948 This prevents bypass of security hardening and performance gains,
2949 especially for curves with specialized EC_METHODs.
2950 By default, if a key encoded with explicit parameters is loaded and later
2951 encoded, the output is still encoded with explicit parameters, even if
2952 internally a "named" EC_GROUP is used for computation.
2956 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2957 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2958 NULL. After this change, only the cofactor parameter can be NULL. It also
2959 does some minimal sanity checks on the passed order.
2964 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2965 An attack is simple, if the first CMS_recipientInfo is valid but the
2966 second CMS_recipientInfo is chosen ciphertext. If the second
2967 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2968 encryption key will be replaced by garbage, and the message cannot be
2969 decoded, but if the RSA decryption fails, the correct encryption key is
2970 used and the recipient will not notice the attack.
2971 As a work around for this potential attack the length of the decrypted
2972 key must be equal to the cipher default key length, in case the
2973 certifiate is not given and all recipientInfo are tried out.
2974 The old behaviour can be re-enabled in the CMS code by setting the
2975 CMS_DEBUG_DECRYPT flag.
2980 * Use Windows installation paths in the mingw builds
2982 Mingw isn't a POSIX environment per se, which means that Windows
2983 paths should be used for installation.
2988 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2990 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2991 This changes the size when using the `genpkey` command when no size is given.
2992 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2993 generation commands to use 2048 bits by default.
2997 * Prevent over long nonces in ChaCha20-Poly1305.
2999 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3000 for every encryption operation. RFC 7539 specifies that the nonce value
3001 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3002 and front pads the nonce with 0 bytes if it is less than 12
3003 bytes. However it also incorrectly allows a nonce to be set of up to 16
3004 bytes. In this case only the last 12 bytes are significant and any
3005 additional leading bytes are ignored.
3007 It is a requirement of using this cipher that nonce values are
3008 unique. Messages encrypted using a reused nonce value are susceptible to
3009 serious confidentiality and integrity attacks. If an application changes
3010 the default nonce length to be longer than 12 bytes and then makes a
3011 change to the leading bytes of the nonce expecting the new value to be a
3012 new unique nonce then such an application could inadvertently encrypt
3013 messages with a reused nonce.
3015 Additionally the ignored bytes in a long nonce are not covered by the
3016 integrity guarantee of this cipher. Any application that relies on the
3017 integrity of these ignored leading bytes of a long nonce may be further
3018 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3019 is safe because no such use sets such a long nonce value. However user
3020 applications that use this cipher directly and set a non-default nonce
3021 length to be longer than 12 bytes may be vulnerable.
3023 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3029 * Added SCA hardening for modular field inversion in EC_GROUP through
3030 a new dedicated field_inv() pointer in EC_METHOD.
3031 This also addresses a leakage affecting conversions from projective
3032 to affine coordinates.
3034 *Billy Bob Brumley, Nicola Tuveri*
3036 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3037 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3041 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3045 * Remove the 'dist' target and add a tarball building script. The
3046 'dist' target has fallen out of use, and it shouldn't be
3047 necessary to configure just to create a source distribution.
3051 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3053 * Timing vulnerability in DSA signature generation
3055 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3056 timing side channel attack. An attacker could use variations in the signing
3057 algorithm to recover the private key.
3059 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3064 * Timing vulnerability in ECDSA signature generation
3066 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3067 timing side channel attack. An attacker could use variations in the signing
3068 algorithm to recover the private key.
3070 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3075 * Add coordinate blinding for EC_POINT and implement projective
3076 coordinate blinding for generic prime curves as a countermeasure to
3077 chosen point SCA attacks.
3079 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3081 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3083 * Client DoS due to large DH parameter
3085 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3086 malicious server can send a very large prime value to the client. This will
3087 cause the client to spend an unreasonably long period of time generating a
3088 key for this prime resulting in a hang until the client has finished. This
3089 could be exploited in a Denial Of Service attack.
3091 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3096 * Cache timing vulnerability in RSA Key Generation
3098 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3099 a cache timing side channel attack. An attacker with sufficient access to
3100 mount cache timing attacks during the RSA key generation process could
3101 recover the private key.
3103 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3104 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3109 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3110 parameter is no longer accepted, as it leads to a corrupt table. NULL
3111 pem_str is reserved for alias entries only.
3115 * Revert blinding in ECDSA sign and instead make problematic addition
3116 length-invariant. Switch even to fixed-length Montgomery multiplication.
3120 * Change generating and checking of primes so that the error rate of not
3121 being prime depends on the intended use based on the size of the input.
3122 For larger primes this will result in more rounds of Miller-Rabin.
3123 The maximal error rate for primes with more than 1080 bits is lowered
3126 *Kurt Roeckx, Annie Yousar*
3128 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3132 * Add blinding to ECDSA and DSA signatures to protect against side channel
3133 attacks discovered by Keegan Ryan (NCC Group).
3137 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3138 now allow empty (zero character) pass phrases.
3142 * Certificate time validation (X509_cmp_time) enforces stricter
3143 compliance with RFC 5280. Fractional seconds and timezone offsets
3144 are no longer allowed.
3148 * Fixed a text canonicalisation bug in CMS
3150 Where a CMS detached signature is used with text content the text goes
3151 through a canonicalisation process first prior to signing or verifying a
3152 signature. This process strips trailing space at the end of lines, converts
3153 line terminators to CRLF and removes additional trailing line terminators
3154 at the end of a file. A bug in the canonicalisation process meant that
3155 some characters, such as form-feed, were incorrectly treated as whitespace
3156 and removed. This is contrary to the specification (RFC5485). This fix
3157 could mean that detached text data signed with an earlier version of
3158 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3159 signed with a fixed OpenSSL may fail to verify with an earlier version of
3160 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3161 and use the "-binary" flag (for the "cms" command line application) or set
3162 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3166 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3168 * Constructed ASN.1 types with a recursive definition could exceed the stack
3170 Constructed ASN.1 types with a recursive definition (such as can be found
3171 in PKCS7) could eventually exceed the stack given malicious input with
3172 excessive recursion. This could result in a Denial Of Service attack. There
3173 are no such structures used within SSL/TLS that come from untrusted sources
3174 so this is considered safe.
3176 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3182 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3184 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3185 effectively reduced to only comparing the least significant bit of each
3186 byte. This allows an attacker to forge messages that would be considered as
3187 authenticated in an amount of tries lower than that guaranteed by the
3188 security claims of the scheme. The module can only be compiled by the
3189 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3191 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3197 * Add a build target 'build_all_generated', to build all generated files
3198 and only that. This can be used to prepare everything that requires
3199 things like perl for a system that lacks perl and then move everything
3200 to that system and do the rest of the build there.
3204 * Backport SSL_OP_NO_RENGOTIATION
3206 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3207 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3208 changes this is no longer possible in 1.1.0. Therefore the new
3209 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3210 1.1.0 to provide equivalent functionality.
3212 Note that if an application built against 1.1.0h headers (or above) is run
3213 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3214 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3218 * Removed the OS390-Unix config target. It relied on a script that doesn't
3223 * rsaz_1024_mul_avx2 overflow bug on x86_64
3225 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3226 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3227 Analysis suggests that attacks against RSA and DSA as a result of this
3228 defect would be very difficult to perform and are not believed likely.
3229 Attacks against DH1024 are considered just feasible, because most of the
3230 work necessary to deduce information about a private key may be performed
3231 offline. The amount of resources required for such an attack would be
3232 significant. However, for an attack on TLS to be meaningful, the server
3233 would have to share the DH1024 private key among multiple clients, which is
3234 no longer an option since CVE-2016-0701.
3236 This only affects processors that support the AVX2 but not ADX extensions
3237 like Intel Haswell (4th generation).
3239 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3240 was originally found via the OSS-Fuzz project.
3245 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3247 * bn_sqrx8x_internal carry bug on x86_64
3249 There is a carry propagating bug in the x86_64 Montgomery squaring
3250 procedure. No EC algorithms are affected. Analysis suggests that attacks
3251 against RSA and DSA as a result of this defect would be very difficult to
3252 perform and are not believed likely. Attacks against DH are considered just
3253 feasible (although very difficult) because most of the work necessary to
3254 deduce information about a private key may be performed offline. The amount
3255 of resources required for such an attack would be very significant and
3256 likely only accessible to a limited number of attackers. An attacker would
3257 additionally need online access to an unpatched system using the target
3258 private key in a scenario with persistent DH parameters and a private
3259 key that is shared between multiple clients.
3261 This only affects processors that support the BMI1, BMI2 and ADX extensions
3262 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3264 This issue was reported to OpenSSL by the OSS-Fuzz project.
3269 * Malformed X.509 IPAddressFamily could cause OOB read
3271 If an X.509 certificate has a malformed IPAddressFamily extension,
3272 OpenSSL could do a one-byte buffer overread. The most likely result
3273 would be an erroneous display of the certificate in text format.
3275 This issue was reported to OpenSSL by the OSS-Fuzz project.
3280 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3282 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3283 platform rather than 'mingw'.
3287 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3288 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3289 which is the minimum version we support.
3293 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3295 * Encrypt-Then-Mac renegotiation crash
3297 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3298 negotiated where it was not in the original handshake (or vice-versa) then
3299 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3300 and servers are affected.
3302 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3307 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3309 * Truncated packet could crash via OOB read
3311 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3312 cipher is being used, then a truncated packet can cause that host to
3313 perform an out-of-bounds read, usually resulting in a crash.
3315 This issue was reported to OpenSSL by Robert Święcki of Google.
3320 * Bad (EC)DHE parameters cause a client crash
3322 If a malicious server supplies bad parameters for a DHE or ECDHE key
3323 exchange then this can result in the client attempting to dereference a
3324 NULL pointer leading to a client crash. This could be exploited in a Denial
3327 This issue was reported to OpenSSL by Guido Vranken.
3332 * BN_mod_exp may produce incorrect results on x86_64
3334 There is a carry propagating bug in the x86_64 Montgomery squaring
3335 procedure. No EC algorithms are affected. Analysis suggests that attacks
3336 against RSA and DSA as a result of this defect would be very difficult to
3337 perform and are not believed likely. Attacks against DH are considered just
3338 feasible (although very difficult) because most of the work necessary to
3339 deduce information about a private key may be performed offline. The amount
3340 of resources required for such an attack would be very significant and
3341 likely only accessible to a limited number of attackers. An attacker would
3342 additionally need online access to an unpatched system using the target
3343 private key in a scenario with persistent DH parameters and a private
3344 key that is shared between multiple clients. For example this can occur by
3345 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3346 similar to CVE-2015-3193 but must be treated as a separate problem.
3348 This issue was reported to OpenSSL by the OSS-Fuzz project.
3353 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3355 * ChaCha20/Poly1305 heap-buffer-overflow
3357 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3358 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3359 crash. This issue is not considered to be exploitable beyond a DoS.
3361 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3366 * CMS Null dereference
3368 Applications parsing invalid CMS structures can crash with a NULL pointer
3369 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3370 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3371 structure callback if an attempt is made to free certain invalid encodings.
3372 Only CHOICE structures using a callback which do not handle NULL value are
3375 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3380 * Montgomery multiplication may produce incorrect results
3382 There is a carry propagating bug in the Broadwell-specific Montgomery
3383 multiplication procedure that handles input lengths divisible by, but
3384 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3385 and DH private keys are impossible. This is because the subroutine in
3386 question is not used in operations with the private key itself and an input
3387 of the attacker's direct choice. Otherwise the bug can manifest itself as
3388 transient authentication and key negotiation failures or reproducible
3389 erroneous outcome of public-key operations with specially crafted input.
3390 Among EC algorithms only Brainpool P-512 curves are affected and one
3391 presumably can attack ECDH key negotiation. Impact was not analyzed in
3392 detail, because pre-requisites for attack are considered unlikely. Namely
3393 multiple clients have to choose the curve in question and the server has to
3394 share the private key among them, neither of which is default behaviour.
3395 Even then only clients that chose the curve will be affected.
3397 This issue was publicly reported as transient failures and was not
3398 initially recognized as a security issue. Thanks to Richard Morgan for
3399 providing reproducible case.
3404 * Removed automatic addition of RPATH in shared libraries and executables,
3405 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3409 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3411 * Fix Use After Free for large message sizes
3413 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3414 message larger than approx 16k is received then the underlying buffer to
3415 store the incoming message is reallocated and moved. Unfortunately a
3416 dangling pointer to the old location is left which results in an attempt to
3417 write to the previously freed location. This is likely to result in a
3418 crash, however it could potentially lead to execution of arbitrary code.
3420 This issue only affects OpenSSL 1.1.0a.
3422 This issue was reported to OpenSSL by Robert Święcki.
3427 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3429 * OCSP Status Request extension unbounded memory growth
3431 A malicious client can send an excessively large OCSP Status Request
3432 extension. If that client continually requests renegotiation, sending a
3433 large OCSP Status Request extension each time, then there will be unbounded
3434 memory growth on the server. This will eventually lead to a Denial Of
3435 Service attack through memory exhaustion. Servers with a default
3436 configuration are vulnerable even if they do not support OCSP. Builds using
3437 the "no-ocsp" build time option are not affected.
3439 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3444 * SSL_peek() hang on empty record
3446 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3447 sends an empty record. This could be exploited by a malicious peer in a
3448 Denial Of Service attack.
3450 This issue was reported to OpenSSL by Alex Gaynor.
3455 * Excessive allocation of memory in tls_get_message_header() and
3456 dtls1_preprocess_fragment()
3458 A (D)TLS message includes 3 bytes for its length in the header for the
3459 message. This would allow for messages up to 16Mb in length. Messages of
3460 this length are excessive and OpenSSL includes a check to ensure that a
3461 peer is sending reasonably sized messages in order to avoid too much memory
3462 being consumed to service a connection. A flaw in the logic of version
3463 1.1.0 means that memory for the message is allocated too early, prior to
3464 the excessive message length check. Due to way memory is allocated in
3465 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3466 to service a connection. This could lead to a Denial of Service through
3467 memory exhaustion. However, the excessive message length check still takes
3468 place, and this would cause the connection to immediately fail. Assuming
3469 that the application calls SSL_free() on the failed connection in a timely
3470 manner then the 21Mb of allocated memory will then be immediately freed
3471 again. Therefore the excessive memory allocation will be transitory in
3472 nature. This then means that there is only a security impact if:
3474 1) The application does not call SSL_free() in a timely manner in the event
3475 that the connection fails
3477 2) The application is working in a constrained environment where there is
3478 very little free memory
3480 3) The attacker initiates multiple connection attempts such that there are
3481 multiple connections in a state where memory has been allocated for the
3482 connection; SSL_free() has not yet been called; and there is insufficient
3483 memory to service the multiple requests.
3485 Except in the instance of (1) above any Denial Of Service is likely to be
3486 transitory because as soon as the connection fails the memory is
3487 subsequently freed again in the SSL_free() call. However there is an
3488 increased risk during this period of application crashes due to the lack of
3489 memory - which would then mean a more serious Denial of Service.
3491 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3492 (CVE-2016-6307 and CVE-2016-6308)
3496 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3497 had to be removed. Primary reason is that vendor assembler can't
3498 assemble our modules with -KPIC flag. As result it, assembly
3499 support, was not even available as option. But its lack means
3500 lack of side-channel resistant code, which is incompatible with
3501 security by todays standards. Fortunately gcc is readily available
3502 prepackaged option, which we firmly point at...
3506 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3508 * Windows command-line tool supports UTF-8 opt-in option for arguments
3509 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3510 (to any value) allows Windows user to access PKCS#12 file generated
3511 with Windows CryptoAPI and protected with non-ASCII password, as well
3512 as files generated under UTF-8 locale on Linux also protected with
3517 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3518 have been disabled by default and removed from DEFAULT, just like RC4.
3519 See the RC4 item below to re-enable both.
3523 * The method for finding the storage location for the Windows RAND seed file
3524 has changed. First we check %RANDFILE%. If that is not set then we check
3525 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3526 all else fails we fall back to C:\.
3530 * The EVP_EncryptUpdate() function has had its return type changed from void
3531 to int. A return of 0 indicates and error while a return of 1 indicates
3536 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3537 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3538 off the constant time implementation for RSA, DSA and DH have been made
3539 no-ops and deprecated.
3543 * Windows RAND implementation was simplified to only get entropy by
3544 calling CryptGenRandom(). Various other RAND-related tickets
3547 *Joseph Wylie Yandle, Rich Salz*
3549 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3550 and `OPENSSL_LH_`, respectively. The old names are available
3551 with API compatibility. They new names are now completely documented.
3555 * Unify TYPE_up_ref(obj) methods signature.
3556 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3557 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3558 int (instead of void) like all others TYPE_up_ref() methods.
3559 So now these methods also check the return value of CRYPTO_atomic_add(),
3560 and the validity of object reference counter.
3562 *fdasilvayy@gmail.com*
3564 * With Windows Visual Studio builds, the .pdb files are installed
3565 alongside the installed libraries and executables. For a static
3566 library installation, ossl_static.pdb is the associate compiler
3567 generated .pdb file to be used when linking programs.
3571 * Remove openssl.spec. Packaging files belong with the packagers.
3575 * Automatic Darwin/OSX configuration has had a refresh, it will now
3576 recognise x86_64 architectures automatically. You can still decide
3577 to build for a different bitness with the environment variable
3578 KERNEL_BITS (can be 32 or 64), for example:
3580 KERNEL_BITS=32 ./config
3584 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3585 256 bit AES and HMAC with SHA256.
3589 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3593 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3597 * To enable users to have their own config files and build file templates,
3598 Configure looks in the directory indicated by the environment variable
3599 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3600 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3601 name and is used as is.
3605 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3606 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3607 X509_CERT_FILE_CTX was removed.
3611 * "shared" builds are now the default. To create only static libraries use
3612 the "no-shared" Configure option.
3616 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3617 All of these option have not worked for some while and are fundamental
3622 * Make various cleanup routines no-ops and mark them as deprecated. Most
3623 global cleanup functions are no longer required because they are handled
3624 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3625 Explicitly de-initing can cause problems (e.g. where a library that uses
3626 OpenSSL de-inits, but an application is still using it). The affected
3627 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3628 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3629 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3630 COMP_zlib_cleanup().
3634 * --strict-warnings no longer enables runtime debugging options
3635 such as REF_DEBUG. Instead, debug options are automatically
3636 enabled with '--debug' builds.
3638 *Andy Polyakov, Emilia Käsper*
3640 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3641 have been moved out of the public header files. New functions for managing
3642 these have been added.
3646 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3647 objects have been moved out of the public header files. New
3648 functions for managing these have been added.
3652 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3653 have been moved out of the public header files. New functions for managing
3654 these have been added.
3658 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3659 moved out of the public header files. New functions for managing these
3664 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3668 * Removed the mk1mf build scripts.
3672 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3673 it is always safe to #include a header now.
3677 * Removed the aged BC-32 config and all its supporting scripts
3681 * Removed support for Ultrix, Netware, and OS/2.
3685 * Add support for HKDF.
3687 *Alessandro Ghedini*
3689 * Add support for blake2b and blake2s
3693 * Added support for "pipelining". Ciphers that have the
3694 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3695 encryptions/decryptions simultaneously. There are currently no built-in
3696 ciphers with this property but the expectation is that engines will be able
3697 to offer it to significantly improve throughput. Support has been extended
3698 into libssl so that multiple records for a single connection can be
3699 processed in one go (for >=TLS 1.1).
3703 * Added the AFALG engine. This is an async capable engine which is able to
3704 offload work to the Linux kernel. In this initial version it only supports
3705 AES128-CBC. The kernel must be version 4.1.0 or greater.
3709 * OpenSSL now uses a new threading API. It is no longer necessary to
3710 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3711 are two supported threading models: pthreads and windows threads. It is
3712 also possible to configure OpenSSL at compile time for "no-threads". The
3713 old threading API should no longer be used. The functions have been
3714 replaced with "no-op" compatibility macros.
3716 *Alessandro Ghedini, Matt Caswell*
3718 * Modify behavior of ALPN to invoke callback after SNI/servername
3719 callback, such that updates to the SSL_CTX affect ALPN.
3723 * Add SSL_CIPHER queries for authentication and key-exchange.
3727 * Changes to the DEFAULT cipherlist:
3728 - Prefer (EC)DHE handshakes over plain RSA.
3729 - Prefer AEAD ciphers over legacy ciphers.
3730 - Prefer ECDSA over RSA when both certificates are available.
3731 - Prefer TLSv1.2 ciphers/PRF.
3732 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3737 * Change the ECC default curve list to be this, in order: x25519,
3738 secp256r1, secp521r1, secp384r1.
3742 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3743 disabled by default. They can be re-enabled using the
3744 enable-weak-ssl-ciphers option to Configure.
3748 * If the server has ALPN configured, but supports no protocols that the
3749 client advertises, send a fatal "no_application_protocol" alert.
3750 This behaviour is SHALL in RFC 7301, though it isn't universally
3751 implemented by other servers.
3755 * Add X25519 support.
3756 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3757 for public and private key encoding using the format documented in
3758 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3759 key generation and key derivation.
3761 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3766 * Deprecate SRP_VBASE_get_by_user.
3767 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3768 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3769 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3770 seed, even if the seed is configured.
3772 Users should use SRP_VBASE_get1_by_user instead. Note that in
3773 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3774 also that even though configuring the SRP seed attempts to hide
3775 invalid usernames by continuing the handshake with fake
3776 credentials, this behaviour is not constant time and no strong
3777 guarantees are made that the handshake is indistinguishable from
3778 that of a valid user.
3782 * Configuration change; it's now possible to build dynamic engines
3783 without having to build shared libraries and vice versa. This
3784 only applies to the engines in `engines/`, those in `crypto/engine/`
3785 will always be built into libcrypto (i.e. "static").
3787 Building dynamic engines is enabled by default; to disable, use
3788 the configuration option "disable-dynamic-engine".
3790 The only requirements for building dynamic engines are the
3791 presence of the DSO module and building with position independent
3792 code, so they will also automatically be disabled if configuring
3793 with "disable-dso" or "disable-pic".
3795 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3796 are also taken away from openssl/opensslconf.h, as they are
3801 * Configuration change; if there is a known flag to compile
3802 position independent code, it will always be applied on the
3803 libcrypto and libssl object files, and never on the application
3804 object files. This means other libraries that use routines from
3805 libcrypto / libssl can be made into shared libraries regardless
3806 of how OpenSSL was configured.
3808 If this isn't desirable, the configuration options "disable-pic"
3809 or "no-pic" can be used to disable the use of PIC. This will
3810 also disable building shared libraries and dynamic engines.
3814 * Removed JPAKE code. It was experimental and has no wide use.
3818 * The INSTALL_PREFIX Makefile variable has been renamed to
3819 DESTDIR. That makes for less confusion on what this variable
3820 is for. Also, the configuration option --install_prefix is
3825 * Heartbeat for TLS has been removed and is disabled by default
3826 for DTLS; configure with enable-heartbeats. Code that uses the
3827 old #define's might need to be updated.
3829 *Emilia Käsper, Rich Salz*
3831 * Rename REF_CHECK to REF_DEBUG.
3835 * New "unified" build system
3837 The "unified" build system is aimed to be a common system for all
3838 platforms we support. With it comes new support for VMS.
3840 This system builds supports building in a different directory tree
3841 than the source tree. It produces one Makefile (for unix family
3842 or lookalikes), or one descrip.mms (for VMS).
3844 The source of information to make the Makefile / descrip.mms is
3845 small files called 'build.info', holding the necessary
3846 information for each directory with source to compile, and a
3847 template in Configurations, like unix-Makefile.tmpl or
3850 With this change, the library names were also renamed on Windows
3851 and on VMS. They now have names that are closer to the standard
3852 on Unix, and include the major version number, and in certain
3853 cases, the architecture they are built for. See "Notes on shared
3854 libraries" in INSTALL.
3856 We rely heavily on the perl module Text::Template.
3860 * Added support for auto-initialisation and de-initialisation of the library.
3861 OpenSSL no longer requires explicit init or deinit routines to be called,
3862 except in certain circumstances. See the OPENSSL_init_crypto() and
3863 OPENSSL_init_ssl() man pages for further information.
3867 * The arguments to the DTLSv1_listen function have changed. Specifically the
3868 "peer" argument is now expected to be a BIO_ADDR object.
3870 * Rewrite of BIO networking library. The BIO library lacked consistent
3871 support of IPv6, and adding it required some more extensive
3872 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3873 which hold all types of addresses and chains of address information.
3874 It also introduces a new API, with functions like BIO_socket,
3875 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3876 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3877 have been adapted accordingly.
3881 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3886 * CRIME protection: disable compression by default, even if OpenSSL is
3887 compiled with zlib enabled. Applications can still enable compression
3888 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3889 using the SSL_CONF library to configure compression.
3893 * The signature of the session callback configured with
3894 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3895 was explicitly marked as `const unsigned char*` instead of
3900 * Always DPURIFY. Remove the use of uninitialized memory in the
3901 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3905 * Removed many obsolete configuration items, including
3906 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3907 MD2_CHAR, MD2_INT, MD2_LONG
3909 IDEA_SHORT, IDEA_LONG
3910 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3912 *Rich Salz, with advice from Andy Polyakov*
3914 * Many BN internals have been moved to an internal header file.
3916 *Rich Salz with help from Andy Polyakov*
3918 * Configuration and writing out the results from it has changed.
3919 Files such as Makefile include/openssl/opensslconf.h and are now
3920 produced through general templates, such as Makefile.in and
3921 crypto/opensslconf.h.in and some help from the perl module
3924 Also, the center of configuration information is no longer
3925 Makefile. Instead, Configure produces a perl module in
3926 configdata.pm which holds most of the config data (in the hash
3927 table %config), the target data that comes from the target
3928 configuration in one of the `Configurations/*.conf` files (in
3933 * To clarify their intended purposes, the Configure options
3934 --prefix and --openssldir change their semantics, and become more
3935 straightforward and less interdependent.
3937 --prefix shall be used exclusively to give the location INSTALLTOP
3938 where programs, scripts, libraries, include files and manuals are
3939 going to be installed. The default is now /usr/local.
3941 --openssldir shall be used exclusively to give the default
3942 location OPENSSLDIR where certificates, private keys, CRLs are
3943 managed. This is also where the default openssl.cnf gets
3945 If the directory given with this option is a relative path, the
3946 values of both the --prefix value and the --openssldir value will
3947 be combined to become OPENSSLDIR.
3948 The default for --openssldir is INSTALLTOP/ssl.
3950 Anyone who uses --openssldir to specify where OpenSSL is to be
3951 installed MUST change to use --prefix instead.
3955 * The GOST engine was out of date and therefore it has been removed. An up
3956 to date GOST engine is now being maintained in an external repository.
3957 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3958 support for GOST ciphersuites (these are only activated if a GOST engine
3963 * EGD is no longer supported by default; use enable-egd when
3966 *Ben Kaduk and Rich Salz*
3968 * The distribution now has Makefile.in files, which are used to
3969 create Makefile's when Configure is run. *Configure must be run
3970 before trying to build now.*
3974 * The return value for SSL_CIPHER_description() for error conditions
3979 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3981 Obtaining and performing DNSSEC validation of TLSA records is
3982 the application's responsibility. The application provides
3983 the TLSA records of its choice to OpenSSL, and these are then
3984 used to authenticate the peer.
3986 The TLSA records need not even come from DNS. They can, for
3987 example, be used to implement local end-entity certificate or
3988 trust-anchor "pinning", where the "pin" data takes the form
3989 of TLSA records, which can augment or replace verification
3990 based on the usual WebPKI public certification authorities.
3994 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3995 continues to support deprecated interfaces in default builds.
3996 However, applications are strongly advised to compile their
3997 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3998 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3999 or the 1.1.0 releases.
4001 In environments in which all applications have been ported to
4002 not use any deprecated interfaces OpenSSL's Configure script
4003 should be used with the --api=1.1.0 option to entirely remove
4004 support for the deprecated features from the library and
4005 unconditionally disable them in the installed headers.
4006 Essentially the same effect can be achieved with the "no-deprecated"
4007 argument to Configure, except that this will always restrict
4008 the build to just the latest API, rather than a fixed API
4011 As applications are ported to future revisions of the API,
4012 they should update their compile-time OPENSSL_API_COMPAT define
4013 accordingly, but in most cases should be able to continue to
4014 compile with later releases.
4016 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4017 0x10000000L and 0x00908000L, respectively. However those
4018 versions did not support the OPENSSL_API_COMPAT feature, and
4019 so applications are not typically tested for explicit support
4020 of just the undeprecated features of either release.
4024 * Add support for setting the minimum and maximum supported protocol.
4025 It can bet set via the SSL_set_min_proto_version() and
4026 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4027 MaxProtocol. It's recommended to use the new APIs to disable
4028 protocols instead of disabling individual protocols using
4029 SSL_set_options() or SSL_CONF's Protocol. This change also
4030 removes support for disabling TLS 1.2 in the OpenSSL TLS
4031 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4035 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4039 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4040 and integrates ECDSA and ECDH functionality into EC. Implementations can
4041 now redirect key generation and no longer need to convert to or from
4044 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4045 include the ec.h header file instead.
4049 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4050 ciphers who are no longer supported and drops support the ephemeral RSA key
4051 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4055 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4056 opaque. For HMAC_CTX, the following constructors and destructors
4059 HMAC_CTX *HMAC_CTX_new(void);
4060 void HMAC_CTX_free(HMAC_CTX *ctx);
4062 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4063 destroy such methods has been added. See EVP_MD_meth_new(3) and
4064 EVP_CIPHER_meth_new(3) for documentation.
4067 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4068 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4069 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4070 an already created structure.
4071 2) For consistency with the majority of our object creators and
4072 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4073 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4074 for deprecated builds.
4078 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4079 cryptographic operations to be performed asynchronously as long as an
4080 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4081 further details. Libssl has also had this capability integrated with the
4082 introduction of the new mode SSL_MODE_ASYNC and associated error
4083 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4084 pages. This work was developed in partnership with Intel Corp.
4088 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4089 always enabled now. If you want to disable the support you should
4090 exclude it using the list of supported ciphers. This also means that the
4091 "-no_ecdhe" option has been removed from s_server.
4095 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4096 SSL_{CTX_}set1_curves() which can set a list.
4100 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4101 curve you want to support using SSL_{CTX_}set1_curves().
4105 * State machine rewrite. The state machine code has been significantly
4106 refactored in order to remove much duplication of code and solve issues
4107 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4108 further details). This change does have some associated API changes.
4109 Notably the SSL_state() function has been removed and replaced by
4110 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4111 SSL_set_state() has been removed altogether. The previous handshake states
4112 defined in ssl.h and ssl3.h have also been removed.
4116 * All instances of the string "ssleay" in the public API were replaced
4117 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4118 Some error codes related to internal RSA_eay API's were renamed.
4122 * The demo files in crypto/threads were moved to demo/threads.
4126 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4129 *Matt Caswell, Rich Salz*
4131 * New ASN.1 embed macro.
4133 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4134 structure is not allocated: it is part of the parent. That is instead of
4142 This reduces memory fragmentation and make it impossible to accidentally
4143 set a mandatory field to NULL.
4145 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4146 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4147 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4152 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4156 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4157 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4158 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4159 DES and RC4 ciphersuites.
4163 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4164 This changes the decoding behaviour for some invalid messages,
4165 though the change is mostly in the more lenient direction, and
4166 legacy behaviour is preserved as much as possible.
4170 * Fix no-stdio build.
4171 *David Woodhouse <David.Woodhouse@intel.com> and also*
4172 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4174 * New testing framework
4175 The testing framework has been largely rewritten and is now using
4176 perl and the perl modules Test::Harness and an extended variant of
4177 Test::More called OpenSSL::Test to do its work. All test scripts in
4178 test/ have been rewritten into test recipes, and all direct calls to
4179 executables in test/Makefile have become individual recipes using the
4180 simplified testing OpenSSL::Test::Simple.
4182 For documentation on our testing modules, do:
4184 perldoc test/testlib/OpenSSL/Test/Simple.pm
4185 perldoc test/testlib/OpenSSL/Test.pm
4189 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4190 are used; the latter aborts on memory leaks (usually checked on exit).
4191 Some undocumented "set malloc, etc., hooks" functions were removed
4192 and others were changed. All are now documented.
4196 * In DSA_generate_parameters_ex, if the provided seed is too short,
4199 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4201 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4202 from RFC4279, RFC4785, RFC5487, RFC5489.
4204 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4205 original RSA_PSK patch.
4209 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4210 era flag was never set throughout the codebase (only read). Also removed
4211 SSL3_FLAGS_POP_BUFFER which was only used if
4212 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4216 * Changed the default name options in the "ca", "crl", "req" and "x509"
4217 to be "oneline" instead of "compat".
4221 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4222 not aware of clients that still exhibit this bug, and the workaround
4223 hasn't been working properly for a while.
4227 * The return type of BIO_number_read() and BIO_number_written() as well as
4228 the corresponding num_read and num_write members in the BIO structure has
4229 changed from unsigned long to uint64_t. On platforms where an unsigned
4230 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4235 * Given the pervasive nature of TLS extensions it is inadvisable to run
4236 OpenSSL without support for them. It also means that maintaining
4237 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4238 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4242 * Removed support for the two export grade static DH ciphersuites
4243 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4244 were newly added (along with a number of other static DH ciphersuites) to
4245 1.0.2. However the two export ones have *never* worked since they were
4246 introduced. It seems strange in any case to be adding new export
4247 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4251 * Version negotiation has been rewritten. In particular SSLv23_method(),
4252 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4253 and turned into macros which simply call the new preferred function names
4254 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4255 should use the new names instead. Also as part of this change the ssl23.h
4256 header file has been removed.
4260 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4261 code and the associated standard is no longer considered fit-for-purpose.
4265 * RT2547 was closed. When generating a private key, try to make the
4266 output file readable only by the owner. This behavior change might
4267 be noticeable when interacting with other software.
4269 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4274 * Added HTTP GET support to the ocsp command.
4278 * Changed default digest for the dgst and enc commands from MD5 to
4283 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4287 * Added support for TLS extended master secret from
4288 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4289 initial patch which was a great help during development.
4293 * All libssl internal structures have been removed from the public header
4294 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4295 now redundant). Users should not attempt to access internal structures
4296 directly. Instead they should use the provided API functions.
4300 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4301 Access to deprecated functions can be re-enabled by running config with
4302 "enable-deprecated". In addition applications wishing to use deprecated
4303 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4304 will, by default, disable some transitive includes that previously existed
4305 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4309 * Added support for OCB mode. OpenSSL has been granted a patent license
4310 compatible with the OpenSSL license for use of OCB. Details are available
4311 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4312 for OCB can be removed by calling config with no-ocb.
4316 * SSLv2 support has been removed. It still supports receiving a SSLv2
4317 compatible client hello.
4321 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4322 done while fixing the error code for the key-too-small case.
4324 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4326 * CA.sh has been removed; use CA.pl instead.
4330 * Removed old DES API.
4334 * Remove various unsupported platforms:
4340 Sinix/ReliantUNIX RM400
4345 16-bit platforms such as WIN16
4349 * Clean up OPENSSL_NO_xxx #define's
4350 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4351 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4352 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4353 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4354 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4355 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4356 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4357 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4358 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4359 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4363 * Cleaned up dead code
4364 Remove all but one '#ifdef undef' which is to be looked at.
4368 * Clean up calling of xxx_free routines.
4369 Just like free(), fix most of the xxx_free routines to accept
4370 NULL. Remove the non-null checks from callers. Save much code.
4374 * Add secure heap for storage of private keys (when possible).
4375 Add BIO_s_secmem(), CBIGNUM, etc.
4376 Contributed by Akamai Technologies under our Corporate CLA.
4380 * Experimental support for a new, fast, unbiased prime candidate generator,
4381 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4383 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4385 * New output format NSS in the sess_id command line tool. This allows
4386 exporting the session id and the master key in NSS keylog format.
4388 *Martin Kaiser <martin@kaiser.cx>*
4390 * Harmonize version and its documentation. -f flag is used to display
4393 *mancha <mancha1@zoho.com>*
4395 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4396 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4398 *mancha <mancha1@zoho.com>*
4400 * Fix some double frees. These are not thought to be exploitable.
4402 *mancha <mancha1@zoho.com>*
4404 * A missing bounds check in the handling of the TLS heartbeat extension
4405 can be used to reveal up to 64k of memory to a connected client or
4408 Thanks for Neel Mehta of Google Security for discovering this bug and to
4409 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4410 preparing the fix ([CVE-2014-0160])
4412 *Adam Langley, Bodo Moeller*
4414 * Fix for the attack described in the paper "Recovering OpenSSL
4415 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4416 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4417 <http://eprint.iacr.org/2014/140>
4419 Thanks to Yuval Yarom and Naomi Benger for discovering this
4420 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4422 *Yuval Yarom and Naomi Benger*
4424 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4425 this fixes a limitation in previous versions of OpenSSL.
4429 * Experimental encrypt-then-mac support.
4431 Experimental support for encrypt then mac from
4432 draft-gutmann-tls-encrypt-then-mac-02.txt
4434 To enable it set the appropriate extension number (0x42 for the test
4435 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4437 For non-compliant peers (i.e. just about everything) this should have no
4440 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4444 * Add EVP support for key wrapping algorithms, to avoid problems with
4445 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4446 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4447 algorithms and include tests cases.
4451 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4456 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4457 MGF1 digest and OAEP label.
4461 * Make openssl verify return errors.
4463 *Chris Palmer <palmer@google.com> and Ben Laurie*
4465 * New function ASN1_TIME_diff to calculate the difference between two
4466 ASN1_TIME structures or one structure and the current time.
4470 * Update fips_test_suite to support multiple command line options. New
4471 test to induce all self test errors in sequence and check expected
4476 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4477 sign or verify all in one operation.
4481 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4482 test programs and fips_test_suite. Includes functionality to parse
4483 the minimal script output of fipsalgest.pl directly.
4487 * Add authorisation parameter to FIPS_module_mode_set().
4491 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4495 * Use separate DRBG fields for internal and external flags. New function
4496 FIPS_drbg_health_check() to perform on demand health checking. Add
4497 generation tests to fips_test_suite with reduced health check interval to
4498 demonstrate periodic health checking. Add "nodh" option to
4499 fips_test_suite to skip very slow DH test.
4503 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4508 * More extensive health check for DRBG checking many more failure modes.
4509 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4510 combination: call this in fips_test_suite.
4514 * Add support for canonical generation of DSA parameter 'g'. See
4517 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4518 POST to handle HMAC cases.
4522 * Add functions FIPS_module_version() and FIPS_module_version_text()
4523 to return numerical and string versions of the FIPS module number.
4527 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4528 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4529 outside the validated module in the FIPS capable OpenSSL.
4533 * Minor change to DRBG entropy callback semantics. In some cases
4534 there is no multiple of the block length between min_len and
4535 max_len. Allow the callback to return more than max_len bytes
4536 of entropy but discard any extra: it is the callback's responsibility
4537 to ensure that the extra data discarded does not impact the
4538 requested amount of entropy.
4542 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4543 information in FIPS186-3, SP800-57 and SP800-131A.
4547 * CCM support via EVP. Interface is very similar to GCM case except we
4548 must supply all data in one chunk (i.e. no update, final) and the
4549 message length must be supplied if AAD is used. Add algorithm test
4554 * Initial version of POST overhaul. Add POST callback to allow the status
4555 of POST to be monitored and/or failures induced. Modify fips_test_suite
4556 to use callback. Always run all selftests even if one fails.
4560 * XTS support including algorithm test driver in the fips_gcmtest program.
4561 Note: this does increase the maximum key length from 32 to 64 bytes but
4562 there should be no binary compatibility issues as existing applications
4563 will never use XTS mode.
4567 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4568 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4569 performs algorithm blocking for unapproved PRNG types. Also do not
4570 set PRNG type in FIPS_mode_set(): leave this to the application.
4571 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4572 the standard OpenSSL PRNG: set additional data to a date time vector.
4576 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4577 This shouldn't present any incompatibility problems because applications
4578 shouldn't be using these directly and any that are will need to rethink
4579 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4583 * Extensive self tests and health checking required by SP800-90 DRBG.
4584 Remove strength parameter from FIPS_drbg_instantiate and always
4585 instantiate at maximum supported strength.
4589 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4593 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4597 * New function DH_compute_key_padded() to compute a DH key and pad with
4598 leading zeroes if needed: this complies with SP800-56A et al.
4602 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4603 anything, incomplete, subject to change and largely untested at present.
4607 * Modify fipscanisteronly build option to only build the necessary object
4608 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4612 * Add experimental option FIPSSYMS to give all symbols in
4613 fipscanister.o and FIPS or fips prefix. This will avoid
4614 conflicts with future versions of OpenSSL. Add perl script
4615 util/fipsas.pl to preprocess assembly language source files
4616 and rename any affected symbols.
4620 * Add selftest checks and algorithm block of non-fips algorithms in
4621 FIPS mode. Remove DES2 from selftests.
4625 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4626 return internal method without any ENGINE dependencies. Add new
4627 tiny fips sign and verify functions.
4631 * New build option no-ec2m to disable characteristic 2 code.
4635 * New build option "fipscanisteronly". This only builds fipscanister.o
4636 and (currently) associated fips utilities. Uses the file Makefile.fips
4637 instead of Makefile.org as the prototype.
4641 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4642 Update fips_gcmtest to use IV generator.
4646 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4647 setting output buffer to NULL. The `*Final` function must be
4648 called although it will not retrieve any additional data. The tag
4649 can be set or retrieved with a ctrl. The IV length is by default 12
4650 bytes (96 bits) but can be set to an alternative value. If the IV
4651 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4656 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4657 underlying do_cipher function handles all cipher semantics itself
4658 including padding and finalisation. This is useful if (for example)
4659 an ENGINE cipher handles block padding itself. The behaviour of
4660 do_cipher is subtly changed if this flag is set: the return value
4661 is the number of characters written to the output buffer (zero is
4662 no longer an error code) or a negative error code. Also if the
4663 input buffer is NULL and length 0 finalisation should be performed.
4667 * If a candidate issuer certificate is already part of the constructed
4668 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4672 * Improve forward-security support: add functions
4674 void SSL_CTX_set_not_resumable_session_callback(
4675 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4676 void SSL_set_not_resumable_session_callback(
4677 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4679 for use by SSL/TLS servers; the callback function will be called whenever a
4680 new session is created, and gets to decide whether the session may be
4681 cached to make it resumable (return 0) or not (return 1). (As by the
4682 SSL/TLS protocol specifications, the session_id sent by the server will be
4683 empty to indicate that the session is not resumable; also, the server will
4684 not generate RFC 4507 (RFC 5077) session tickets.)
4686 A simple reasonable callback implementation is to return is_forward_secure.
4687 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4688 by the SSL/TLS server library, indicating whether it can provide forward
4691 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4693 * New -verify_name option in command line utilities to set verification
4698 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4699 Add CMAC pkey methods.
4703 * Experimental renegotiation in s_server -www mode. If the client
4704 browses /reneg connection is renegotiated. If /renegcert it is
4705 renegotiated requesting a certificate.
4709 * Add an "external" session cache for debugging purposes to s_server. This
4710 should help trace issues which normally are only apparent in deployed
4711 multi-process servers.
4715 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4716 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4717 BIO_set_cipher() and some obscure PEM functions were changed so they
4718 can now return an error. The RAND changes required a change to the
4719 RAND_METHOD structure.
4723 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4724 a gcc attribute to warn if the result of a function is ignored. This
4725 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4726 whose return value is often ignored.
4730 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4731 These allow SCTs (signed certificate timestamps) to be requested and
4732 validated when establishing a connection.
4734 *Rob Percival <robpercival@google.com>*
4739 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4741 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4742 used even when parsing explicit parameters, when loading a encoded key
4743 or calling `EC_GROUP_new_from_ecpkparameters()`/
4744 `EC_GROUP_new_from_ecparameters()`.
4745 This prevents bypass of security hardening and performance gains,
4746 especially for curves with specialized EC_METHODs.
4747 By default, if a key encoded with explicit parameters is loaded and later
4748 encoded, the output is still encoded with explicit parameters, even if
4749 internally a "named" EC_GROUP is used for computation.
4753 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4754 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4755 NULL. After this change, only the cofactor parameter can be NULL. It also
4756 does some minimal sanity checks on the passed order.
4761 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4762 An attack is simple, if the first CMS_recipientInfo is valid but the
4763 second CMS_recipientInfo is chosen ciphertext. If the second
4764 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4765 encryption key will be replaced by garbage, and the message cannot be
4766 decoded, but if the RSA decryption fails, the correct encryption key is
4767 used and the recipient will not notice the attack.
4768 As a work around for this potential attack the length of the decrypted
4769 key must be equal to the cipher default key length, in case the
4770 certifiate is not given and all recipientInfo are tried out.
4771 The old behaviour can be re-enabled in the CMS code by setting the
4772 CMS_DEBUG_DECRYPT flag.
4777 * Document issue with installation paths in diverse Windows builds
4779 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4780 binaries and run-time config file.
4785 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4787 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4788 This changes the size when using the `genpkey` command when no size is given.
4789 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4790 generation commands to use 2048 bits by default.
4794 * Add FIPS support for Android Arm 64-bit
4796 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4797 Module in Version 2.0.10. For some reason, the corresponding target
4798 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4799 built with FIPS support on Android Arm 64-bit. This omission has been
4802 *Matthias St. Pierre*
4804 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4806 * 0-byte record padding oracle
4808 If an application encounters a fatal protocol error and then calls
4809 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4810 then OpenSSL can respond differently to the calling application if a 0 byte
4811 record is received with invalid padding compared to if a 0 byte record is
4812 received with an invalid MAC. If the application then behaves differently
4813 based on that in a way that is detectable to the remote peer, then this
4814 amounts to a padding oracle that could be used to decrypt data.
4816 In order for this to be exploitable "non-stitched" ciphersuites must be in
4817 use. Stitched ciphersuites are optimised implementations of certain
4818 commonly used ciphersuites. Also the application must call SSL_shutdown()
4819 twice even if a protocol error has occurred (applications should not do
4820 this but some do anyway).
4822 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4823 Aviram, with additional investigation by Steven Collison and Andrew
4824 Hourselt. It was reported to OpenSSL on 10th December 2018.
4829 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4833 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4835 * Microarchitecture timing vulnerability in ECC scalar multiplication
4837 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4838 shown to be vulnerable to a microarchitecture timing side channel attack.
4839 An attacker with sufficient access to mount local timing attacks during
4840 ECDSA signature generation could recover the private key.
4842 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4843 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4849 * Timing vulnerability in DSA signature generation
4851 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4852 timing side channel attack. An attacker could use variations in the signing
4853 algorithm to recover the private key.
4855 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4860 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4861 Module, accidentally introduced while backporting security fixes from the
4862 development branch and hindering the use of ECC in FIPS mode.
4866 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4868 * Client DoS due to large DH parameter
4870 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4871 malicious server can send a very large prime value to the client. This will
4872 cause the client to spend an unreasonably long period of time generating a
4873 key for this prime resulting in a hang until the client has finished. This
4874 could be exploited in a Denial Of Service attack.
4876 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4881 * Cache timing vulnerability in RSA Key Generation
4883 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4884 a cache timing side channel attack. An attacker with sufficient access to
4885 mount cache timing attacks during the RSA key generation process could
4886 recover the private key.
4888 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4889 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4894 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4895 parameter is no longer accepted, as it leads to a corrupt table. NULL
4896 pem_str is reserved for alias entries only.
4900 * Revert blinding in ECDSA sign and instead make problematic addition
4901 length-invariant. Switch even to fixed-length Montgomery multiplication.
4905 * Change generating and checking of primes so that the error rate of not
4906 being prime depends on the intended use based on the size of the input.
4907 For larger primes this will result in more rounds of Miller-Rabin.
4908 The maximal error rate for primes with more than 1080 bits is lowered
4911 *Kurt Roeckx, Annie Yousar*
4913 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4917 * Add blinding to ECDSA and DSA signatures to protect against side channel
4918 attacks discovered by Keegan Ryan (NCC Group).
4922 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4923 now allow empty (zero character) pass phrases.
4927 * Certificate time validation (X509_cmp_time) enforces stricter
4928 compliance with RFC 5280. Fractional seconds and timezone offsets
4929 are no longer allowed.
4933 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4935 * Constructed ASN.1 types with a recursive definition could exceed the stack
4937 Constructed ASN.1 types with a recursive definition (such as can be found
4938 in PKCS7) could eventually exceed the stack given malicious input with
4939 excessive recursion. This could result in a Denial Of Service attack. There
4940 are no such structures used within SSL/TLS that come from untrusted sources
4941 so this is considered safe.
4943 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4949 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4951 * Read/write after SSL object in error state
4953 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4954 mechanism. The intent was that if a fatal error occurred during a handshake
4955 then OpenSSL would move into the error state and would immediately fail if
4956 you attempted to continue the handshake. This works as designed for the
4957 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4958 SSL_connect()), however due to a bug it does not work correctly if
4959 SSL_read() or SSL_write() is called directly. In that scenario, if the
4960 handshake fails then a fatal error will be returned in the initial function
4961 call. If SSL_read()/SSL_write() is subsequently called by the application
4962 for the same SSL object then it will succeed and the data is passed without
4963 being decrypted/encrypted directly from the SSL/TLS record layer.
4965 In order to exploit this issue an application bug would have to be present
4966 that resulted in a call to SSL_read()/SSL_write() being issued after having
4967 already received a fatal error.
4969 This issue was reported to OpenSSL by David Benjamin (Google).
4974 * rsaz_1024_mul_avx2 overflow bug on x86_64
4976 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4977 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4978 Analysis suggests that attacks against RSA and DSA as a result of this
4979 defect would be very difficult to perform and are not believed likely.
4980 Attacks against DH1024 are considered just feasible, because most of the
4981 work necessary to deduce information about a private key may be performed
4982 offline. The amount of resources required for such an attack would be
4983 significant. However, for an attack on TLS to be meaningful, the server
4984 would have to share the DH1024 private key among multiple clients, which is
4985 no longer an option since CVE-2016-0701.
4987 This only affects processors that support the AVX2 but not ADX extensions
4988 like Intel Haswell (4th generation).
4990 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4991 was originally found via the OSS-Fuzz project.
4996 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4998 * bn_sqrx8x_internal carry bug on x86_64
5000 There is a carry propagating bug in the x86_64 Montgomery squaring
5001 procedure. No EC algorithms are affected. Analysis suggests that attacks
5002 against RSA and DSA as a result of this defect would be very difficult to
5003 perform and are not believed likely. Attacks against DH are considered just
5004 feasible (although very difficult) because most of the work necessary to
5005 deduce information about a private key may be performed offline. The amount
5006 of resources required for such an attack would be very significant and
5007 likely only accessible to a limited number of attackers. An attacker would
5008 additionally need online access to an unpatched system using the target
5009 private key in a scenario with persistent DH parameters and a private
5010 key that is shared between multiple clients.
5012 This only affects processors that support the BMI1, BMI2 and ADX extensions
5013 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5015 This issue was reported to OpenSSL by the OSS-Fuzz project.
5020 * Malformed X.509 IPAddressFamily could cause OOB read
5022 If an X.509 certificate has a malformed IPAddressFamily extension,
5023 OpenSSL could do a one-byte buffer overread. The most likely result
5024 would be an erroneous display of the certificate in text format.
5026 This issue was reported to OpenSSL by the OSS-Fuzz project.
5030 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5032 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5033 platform rather than 'mingw'.
5037 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5039 * Truncated packet could crash via OOB read
5041 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5042 cipher is being used, then a truncated packet can cause that host to
5043 perform an out-of-bounds read, usually resulting in a crash.
5045 This issue was reported to OpenSSL by Robert Święcki of Google.
5050 * BN_mod_exp may produce incorrect results on x86_64
5052 There is a carry propagating bug in the x86_64 Montgomery squaring
5053 procedure. No EC algorithms are affected. Analysis suggests that attacks
5054 against RSA and DSA as a result of this defect would be very difficult to
5055 perform and are not believed likely. Attacks against DH are considered just
5056 feasible (although very difficult) because most of the work necessary to
5057 deduce information about a private key may be performed offline. The amount
5058 of resources required for such an attack would be very significant and
5059 likely only accessible to a limited number of attackers. An attacker would
5060 additionally need online access to an unpatched system using the target
5061 private key in a scenario with persistent DH parameters and a private
5062 key that is shared between multiple clients. For example this can occur by
5063 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5064 similar to CVE-2015-3193 but must be treated as a separate problem.
5066 This issue was reported to OpenSSL by the OSS-Fuzz project.
5071 * Montgomery multiplication may produce incorrect results
5073 There is a carry propagating bug in the Broadwell-specific Montgomery
5074 multiplication procedure that handles input lengths divisible by, but
5075 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5076 and DH private keys are impossible. This is because the subroutine in
5077 question is not used in operations with the private key itself and an input
5078 of the attacker's direct choice. Otherwise the bug can manifest itself as
5079 transient authentication and key negotiation failures or reproducible
5080 erroneous outcome of public-key operations with specially crafted input.
5081 Among EC algorithms only Brainpool P-512 curves are affected and one
5082 presumably can attack ECDH key negotiation. Impact was not analyzed in
5083 detail, because pre-requisites for attack are considered unlikely. Namely
5084 multiple clients have to choose the curve in question and the server has to
5085 share the private key among them, neither of which is default behaviour.
5086 Even then only clients that chose the curve will be affected.
5088 This issue was publicly reported as transient failures and was not
5089 initially recognized as a security issue. Thanks to Richard Morgan for
5090 providing reproducible case.
5095 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5096 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5097 prevent issues where no progress is being made and the peer continually
5098 sends unrecognised record types, using up resources processing them.
5102 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5104 * Missing CRL sanity check
5106 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5107 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5108 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5110 This issue only affects the OpenSSL 1.0.2i
5115 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5117 * OCSP Status Request extension unbounded memory growth
5119 A malicious client can send an excessively large OCSP Status Request
5120 extension. If that client continually requests renegotiation, sending a
5121 large OCSP Status Request extension each time, then there will be unbounded
5122 memory growth on the server. This will eventually lead to a Denial Of
5123 Service attack through memory exhaustion. Servers with a default
5124 configuration are vulnerable even if they do not support OCSP. Builds using
5125 the "no-ocsp" build time option are not affected.
5127 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5132 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5135 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5141 * OOB write in MDC2_Update()
5143 An overflow can occur in MDC2_Update() either if called directly or
5144 through the EVP_DigestUpdate() function using MDC2. If an attacker
5145 is able to supply very large amounts of input data after a previous
5146 call to EVP_EncryptUpdate() with a partial block then a length check
5147 can overflow resulting in a heap corruption.
5149 The amount of data needed is comparable to SIZE_MAX which is impractical
5152 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5157 * Malformed SHA512 ticket DoS
5159 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5160 DoS attack where a malformed ticket will result in an OOB read which will
5163 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5164 a custom server callback and ticket lookup mechanism.
5166 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5171 * OOB write in BN_bn2dec()
5173 The function BN_bn2dec() does not check the return value of BN_div_word().
5174 This can cause an OOB write if an application uses this function with an
5175 overly large BIGNUM. This could be a problem if an overly large certificate
5176 or CRL is printed out from an untrusted source. TLS is not affected because
5177 record limits will reject an oversized certificate before it is parsed.
5179 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5184 * OOB read in TS_OBJ_print_bio()
5186 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5187 the total length the OID text representation would use and not the amount
5188 of data written. This will result in OOB reads when large OIDs are
5191 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5196 * Pointer arithmetic undefined behaviour
5198 Avoid some undefined pointer arithmetic
5200 A common idiom in the codebase is to check limits in the following manner:
5203 Where "p" points to some malloc'd data of SIZE bytes and
5206 "len" here could be from some externally supplied data (e.g. from a TLS
5209 The rules of C pointer arithmetic are such that "p + len" is only well
5210 defined where len <= SIZE. Therefore the above idiom is actually
5211 undefined behaviour.
5213 For example this could cause problems if some malloc implementation
5214 provides an address for "p" such that "p + len" actually overflows for
5215 values of len that are too big and therefore p + len < limit.
5217 This issue was reported to OpenSSL by Guido Vranken
5222 * Constant time flag not preserved in DSA signing
5224 Operations in the DSA signing algorithm should run in constant time in
5225 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5226 implementation means that a non-constant time codepath is followed for
5227 certain operations. This has been demonstrated through a cache-timing
5228 attack to be sufficient for an attacker to recover the private DSA key.
5230 This issue was reported by César Pereida (Aalto University), Billy Brumley
5231 (Tampere University of Technology), and Yuval Yarom (The University of
5232 Adelaide and NICTA).
5237 * DTLS buffered message DoS
5239 In a DTLS connection where handshake messages are delivered out-of-order
5240 those messages that OpenSSL is not yet ready to process will be buffered
5241 for later use. Under certain circumstances, a flaw in the logic means that
5242 those messages do not get removed from the buffer even though the handshake
5243 has been completed. An attacker could force up to approx. 15 messages to
5244 remain in the buffer when they are no longer required. These messages will
5245 be cleared when the DTLS connection is closed. The default maximum size for
5246 a message is 100k. Therefore the attacker could force an additional 1500k
5247 to be consumed per connection. By opening many simulataneous connections an
5248 attacker could cause a DoS attack through memory exhaustion.
5250 This issue was reported to OpenSSL by Quan Luo.
5255 * DTLS replay protection DoS
5257 A flaw in the DTLS replay attack protection mechanism means that records
5258 that arrive for future epochs update the replay protection "window" before
5259 the MAC for the record has been validated. This could be exploited by an
5260 attacker by sending a record for the next epoch (which does not have to
5261 decrypt or have a valid MAC), with a very large sequence number. This means
5262 that all subsequent legitimate packets are dropped causing a denial of
5263 service for a specific DTLS connection.
5265 This issue was reported to OpenSSL by the OCAP audit team.
5270 * Certificate message OOB reads
5272 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5273 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5274 theoretical DoS risk but this has not been observed in practice on common
5277 The messages affected are client certificate, client certificate request
5278 and server certificate. As a result the attack can only be performed
5279 against a client or a server which enables client authentication.
5281 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5286 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5288 * Prevent padding oracle in AES-NI CBC MAC check
5290 A MITM attacker can use a padding oracle attack to decrypt traffic
5291 when the connection uses an AES CBC cipher and the server support
5294 This issue was introduced as part of the fix for Lucky 13 padding
5295 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5296 constant time by making sure that always the same bytes are read and
5297 compared against either the MAC or padding bytes. But it no longer
5298 checked that there was enough data to have both the MAC and padding
5301 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5305 * Fix EVP_EncodeUpdate overflow
5307 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5308 Base64 encoding of binary data. If an attacker is able to supply very large
5309 amounts of input data then a length check can overflow resulting in a heap
5312 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5313 the `PEM_write_bio*` family of functions. These are mainly used within the
5314 OpenSSL command line applications, so any application which processes data
5315 from an untrusted source and outputs it as a PEM file should be considered
5316 vulnerable to this issue. User applications that call these APIs directly
5317 with large amounts of untrusted data may also be vulnerable.
5319 This issue was reported by Guido Vranken.
5324 * Fix EVP_EncryptUpdate overflow
5326 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5327 is able to supply very large amounts of input data after a previous call to
5328 EVP_EncryptUpdate() with a partial block then a length check can overflow
5329 resulting in a heap corruption. Following an analysis of all OpenSSL
5330 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5331 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5332 the first called function after an EVP_EncryptInit(), and therefore that
5333 specific call must be safe. The second form is where the length passed to
5334 EVP_EncryptUpdate() can be seen from the code to be some small value and
5335 therefore there is no possibility of an overflow. Since all instances are
5336 one of these two forms, it is believed that there can be no overflows in
5337 internal code due to this problem. It should be noted that
5338 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5339 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5340 of these calls have also been analysed too and it is believed there are no
5341 instances in internal usage where an overflow could occur.
5343 This issue was reported by Guido Vranken.
5348 * Prevent ASN.1 BIO excessive memory allocation
5350 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5351 a short invalid encoding can cause allocation of large amounts of memory
5352 potentially consuming excessive resources or exhausting memory.
5354 Any application parsing untrusted data through d2i BIO functions is
5355 affected. The memory based functions such as d2i_X509() are *not* affected.
5356 Since the memory based functions are used by the TLS library, TLS
5357 applications are not affected.
5359 This issue was reported by Brian Carpenter.
5366 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5367 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5368 in arbitrary stack data being returned in the buffer.
5370 This issue was reported by Guido Vranken.
5375 * Modify behavior of ALPN to invoke callback after SNI/servername
5376 callback, such that updates to the SSL_CTX affect ALPN.
5380 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5385 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5386 methods are enabled and ssl2 is disabled the methods return NULL.
5390 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5392 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5393 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5394 provide any "EXPORT" or "LOW" strength ciphers.
5398 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5399 is by default disabled at build-time. Builds that are not configured with
5400 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5401 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5402 will need to explicitly call either of:
5404 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5406 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5408 as appropriate. Even if either of those is used, or the application
5409 explicitly uses the version-specific SSLv2_method() or its client and
5410 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5411 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5412 ciphers, and SSLv2 56-bit DES are no longer available.
5417 * Fix a double-free in DSA code
5419 A double free bug was discovered when OpenSSL parses malformed DSA private
5420 keys and could lead to a DoS attack or memory corruption for applications
5421 that receive DSA private keys from untrusted sources. This scenario is
5424 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5430 * Disable SRP fake user seed to address a server memory leak.
5432 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5434 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5435 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5436 was changed to ignore the "fake user" SRP seed, even if the seed
5439 Users should use SRP_VBASE_get1_by_user instead. Note that in
5440 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5441 also that even though configuring the SRP seed attempts to hide
5442 invalid usernames by continuing the handshake with fake
5443 credentials, this behaviour is not constant time and no strong
5444 guarantees are made that the handshake is indistinguishable from
5445 that of a valid user.
5450 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5452 In the BN_hex2bn function the number of hex digits is calculated using an
5453 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5454 large values of `i` this can result in `bn_expand` not allocating any
5455 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5456 field as NULL leading to a subsequent NULL ptr deref. For very large values
5457 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5458 In this case memory is allocated to the internal BIGNUM data field, but it
5459 is insufficiently sized leading to heap corruption. A similar issue exists
5460 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5461 is ever called by user applications with very large untrusted hex/dec data.
5462 This is anticipated to be a rare occurrence.
5464 All OpenSSL internal usage of these functions use data that is not expected
5465 to be untrusted, e.g. config file data or application command line
5466 arguments. If user developed applications generate config file data based
5467 on untrusted data then it is possible that this could also lead to security
5468 consequences. This is also anticipated to be rare.
5470 This issue was reported to OpenSSL by Guido Vranken.
5475 * Fix memory issues in `BIO_*printf` functions
5477 The internal `fmtstr` function used in processing a "%s" format string in
5478 the `BIO_*printf` functions could overflow while calculating the length of a
5479 string and cause an OOB read when printing very long strings.
5481 Additionally the internal `doapr_outch` function can attempt to write to an
5482 OOB memory location (at an offset from the NULL pointer) in the event of a
5483 memory allocation failure. In 1.0.2 and below this could be caused where
5484 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5485 could be in processing a very long "%s" format string. Memory leaks can
5488 The first issue may mask the second issue dependent on compiler behaviour.
5489 These problems could enable attacks where large amounts of untrusted data
5490 is passed to the `BIO_*printf` functions. If applications use these functions
5491 in this way then they could be vulnerable. OpenSSL itself uses these
5492 functions when printing out human-readable dumps of ASN.1 data. Therefore
5493 applications that print this data could be vulnerable if the data is from
5494 untrusted sources. OpenSSL command line applications could also be
5495 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5496 as command line arguments.
5498 Libssl is not considered directly vulnerable. Additionally certificates etc
5499 received via remote connections via libssl are also unlikely to be able to
5500 trigger these issues because of message size limits enforced within libssl.
5502 This issue was reported to OpenSSL Guido Vranken.
5507 * Side channel attack on modular exponentiation
5509 A side-channel attack was found which makes use of cache-bank conflicts on
5510 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5511 of RSA keys. The ability to exploit this issue is limited as it relies on
5512 an attacker who has control of code in a thread running on the same
5513 hyper-threaded core as the victim thread which is performing decryptions.
5515 This issue was reported to OpenSSL by Yuval Yarom, The University of
5516 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5517 Nadia Heninger, University of Pennsylvania with more information at
5518 <http://cachebleed.info>.
5523 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5524 if no keysize is specified with default_bits. This fixes an
5525 omission in an earlier change that changed all RSA/DSA key generation
5526 commands to use 2048 bits by default.
5530 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5532 * DH small subgroups
5534 Historically OpenSSL only ever generated DH parameters based on "safe"
5535 primes. More recently (in version 1.0.2) support was provided for
5536 generating X9.42 style parameter files such as those required for RFC 5114
5537 support. The primes used in such files may not be "safe". Where an
5538 application is using DH configured with parameters based on primes that are
5539 not "safe" then an attacker could use this fact to find a peer's private
5540 DH exponent. This attack requires that the attacker complete multiple
5541 handshakes in which the peer uses the same private DH exponent. For example
5542 this could be used to discover a TLS server's private DH exponent if it's
5543 reusing the private DH exponent or it's using a static DH ciphersuite.
5545 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5546 TLS. It is not on by default. If the option is not set then the server
5547 reuses the same private DH exponent for the life of the server process and
5548 would be vulnerable to this attack. It is believed that many popular
5549 applications do set this option and would therefore not be at risk.
5551 The fix for this issue adds an additional check where a "q" parameter is
5552 available (as is the case in X9.42 based parameters). This detects the
5553 only known attack, and is the only possible defense for static DH
5554 ciphersuites. This could have some performance impact.
5556 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5557 default and cannot be disabled. This could have some performance impact.
5559 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5564 * SSLv2 doesn't block disabled ciphers
5566 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5567 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5568 been disabled, provided that the SSLv2 protocol was not also disabled via
5571 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5572 and Sebastian Schinzel.
5577 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5579 * BN_mod_exp may produce incorrect results on x86_64
5581 There is a carry propagating bug in the x86_64 Montgomery squaring
5582 procedure. No EC algorithms are affected. Analysis suggests that attacks
5583 against RSA and DSA as a result of this defect would be very difficult to
5584 perform and are not believed likely. Attacks against DH are considered just
5585 feasible (although very difficult) because most of the work necessary to
5586 deduce information about a private key may be performed offline. The amount
5587 of resources required for such an attack would be very significant and
5588 likely only accessible to a limited number of attackers. An attacker would
5589 additionally need online access to an unpatched system using the target
5590 private key in a scenario with persistent DH parameters and a private
5591 key that is shared between multiple clients. For example this can occur by
5592 default in OpenSSL DHE based SSL/TLS ciphersuites.
5594 This issue was reported to OpenSSL by Hanno Böck.
5599 * Certificate verify crash with missing PSS parameter
5601 The signature verification routines will crash with a NULL pointer
5602 dereference if presented with an ASN.1 signature using the RSA PSS
5603 algorithm and absent mask generation function parameter. Since these
5604 routines are used to verify certificate signature algorithms this can be
5605 used to crash any certificate verification operation and exploited in a
5606 DoS attack. Any application which performs certificate verification is
5607 vulnerable including OpenSSL clients and servers which enable client
5610 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5615 * X509_ATTRIBUTE memory leak
5617 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5618 memory. This structure is used by the PKCS#7 and CMS routines so any
5619 application which reads PKCS#7 or CMS data from untrusted sources is
5620 affected. SSL/TLS is not affected.
5622 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5628 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5629 This changes the decoding behaviour for some invalid messages,
5630 though the change is mostly in the more lenient direction, and
5631 legacy behaviour is preserved as much as possible.
5635 * In DSA_generate_parameters_ex, if the provided seed is too short,
5638 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5640 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5642 * Alternate chains certificate forgery
5644 During certificate verification, OpenSSL will attempt to find an
5645 alternative certificate chain if the first attempt to build such a chain
5646 fails. An error in the implementation of this logic can mean that an
5647 attacker could cause certain checks on untrusted certificates to be
5648 bypassed, such as the CA flag, enabling them to use a valid leaf
5649 certificate to act as a CA and "issue" an invalid certificate.
5651 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5656 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5658 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5659 incompatibility in the handling of HMAC. The previous ABI has now been
5664 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5666 * Malformed ECParameters causes infinite loop
5668 When processing an ECParameters structure OpenSSL enters an infinite loop
5669 if the curve specified is over a specially malformed binary polynomial
5672 This can be used to perform denial of service against any
5673 system which processes public keys, certificate requests or
5674 certificates. This includes TLS clients and TLS servers with
5675 client authentication enabled.
5677 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5682 * Exploitable out-of-bounds read in X509_cmp_time
5684 X509_cmp_time does not properly check the length of the ASN1_TIME
5685 string and can read a few bytes out of bounds. In addition,
5686 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5689 An attacker can use this to craft malformed certificates and CRLs of
5690 various sizes and potentially cause a segmentation fault, resulting in
5691 a DoS on applications that verify certificates or CRLs. TLS clients
5692 that verify CRLs are affected. TLS clients and servers with client
5693 authentication enabled may be affected if they use custom verification
5696 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5697 independently by Hanno Böck.
5702 * PKCS7 crash with missing EnvelopedContent
5704 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5705 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5706 with missing content and trigger a NULL pointer dereference on parsing.
5708 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5709 structures from untrusted sources are affected. OpenSSL clients and
5710 servers are not affected.
5712 This issue was reported to OpenSSL by Michal Zalewski (Google).
5717 * CMS verify infinite loop with unknown hash function
5719 When verifying a signedData message the CMS code can enter an infinite loop
5720 if presented with an unknown hash function OID. This can be used to perform
5721 denial of service against any system which verifies signedData messages using
5723 This issue was reported to OpenSSL by Johannes Bauer.
5728 * Race condition handling NewSessionTicket
5730 If a NewSessionTicket is received by a multi-threaded client when attempting to
5731 reuse a previous ticket then a race condition can occur potentially leading to
5732 a double free of the ticket data.
5737 * Only support 256-bit or stronger elliptic curves with the
5738 'ecdh_auto' setting (server) or by default (client). Of supported
5739 curves, prefer P-256 (both).
5743 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5745 * ClientHello sigalgs DoS fix
5747 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5748 invalid signature algorithms extension a NULL pointer dereference will
5749 occur. This can be exploited in a DoS attack against the server.
5751 This issue was was reported to OpenSSL by David Ramos of Stanford
5755 *Stephen Henson and Matt Caswell*
5757 * Multiblock corrupted pointer fix
5759 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5760 feature only applies on 64 bit x86 architecture platforms that support AES
5761 NI instructions. A defect in the implementation of "multiblock" can cause
5762 OpenSSL's internal write buffer to become incorrectly set to NULL when
5763 using non-blocking IO. Typically, when the user application is using a
5764 socket BIO for writing, this will only result in a failed connection.
5765 However if some other BIO is used then it is likely that a segmentation
5766 fault will be triggered, thus enabling a potential DoS attack.
5768 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5773 * Segmentation fault in DTLSv1_listen fix
5775 The DTLSv1_listen function is intended to be stateless and processes the
5776 initial ClientHello from many peers. It is common for user code to loop
5777 over the call to DTLSv1_listen until a valid ClientHello is received with
5778 an associated cookie. A defect in the implementation of DTLSv1_listen means
5779 that state is preserved in the SSL object from one invocation to the next
5780 that can lead to a segmentation fault. Errors processing the initial
5781 ClientHello can trigger this scenario. An example of such an error could be
5782 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5785 This issue was reported to OpenSSL by Per Allansson.
5790 * Segmentation fault in ASN1_TYPE_cmp fix
5792 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5793 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5794 certificate signature algorithm consistency this can be used to crash any
5795 certificate verification operation and exploited in a DoS attack. Any
5796 application which performs certificate verification is vulnerable including
5797 OpenSSL clients and servers which enable client authentication.
5802 * Segmentation fault for invalid PSS parameters fix
5804 The signature verification routines will crash with a NULL pointer
5805 dereference if presented with an ASN.1 signature using the RSA PSS
5806 algorithm and invalid parameters. Since these routines are used to verify
5807 certificate signature algorithms this can be used to crash any
5808 certificate verification operation and exploited in a DoS attack. Any
5809 application which performs certificate verification is vulnerable including
5810 OpenSSL clients and servers which enable client authentication.
5812 This issue was was reported to OpenSSL by Brian Carpenter.
5817 * ASN.1 structure reuse memory corruption fix
5819 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5820 memory corruption via an invalid write. Such reuse is and has been
5821 strongly discouraged and is believed to be rare.
5823 Applications that parse structures containing CHOICE or ANY DEFINED BY
5824 components may be affected. Certificate parsing (d2i_X509 and related
5825 functions) are however not affected. OpenSSL clients and servers are
5831 * PKCS7 NULL pointer dereferences fix
5833 The PKCS#7 parsing code does not handle missing outer ContentInfo
5834 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5835 missing content and trigger a NULL pointer dereference on parsing.
5837 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5838 otherwise parse PKCS#7 structures from untrusted sources are
5839 affected. OpenSSL clients and servers are not affected.
5841 This issue was reported to OpenSSL by Michal Zalewski (Google).
5846 * DoS via reachable assert in SSLv2 servers fix
5848 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5849 servers that both support SSLv2 and enable export cipher suites by sending
5850 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5852 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5853 (OpenSSL development team).
5858 * Empty CKE with client auth and DHE fix
5860 If client auth is used then a server can seg fault in the event of a DHE
5861 ciphersuite being selected and a zero length ClientKeyExchange message
5862 being sent by the client. This could be exploited in a DoS attack.
5867 * Handshake with unseeded PRNG fix
5869 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5870 with an unseeded PRNG. The conditions are:
5871 - The client is on a platform where the PRNG has not been seeded
5872 automatically, and the user has not seeded manually
5873 - A protocol specific client method version has been used (i.e. not
5874 SSL_client_methodv23)
5875 - A ciphersuite is used that does not require additional random data from
5876 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5878 If the handshake succeeds then the client random that has been used will
5879 have been generated from a PRNG with insufficient entropy and therefore the
5880 output may be predictable.
5882 For example using the following command with an unseeded openssl will
5883 succeed on an unpatched platform:
5885 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5890 * Use After Free following d2i_ECPrivatekey error fix
5892 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5893 could cause a use after free condition. This, in turn, could cause a double
5894 free in several private key parsing functions (such as d2i_PrivateKey
5895 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5896 for applications that receive EC private keys from untrusted
5897 sources. This scenario is considered rare.
5899 This issue was discovered by the BoringSSL project and fixed in their
5905 * X509_to_X509_REQ NULL pointer deref fix
5907 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5908 the certificate key is invalid. This function is rarely used in practice.
5910 This issue was discovered by Brian Carpenter.
5915 * Removed the export ciphers from the DEFAULT ciphers
5919 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5921 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5922 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5923 So far those who have to target multiple platforms would compromise
5924 and argue that binary targeting say ARMv5 would still execute on
5925 ARMv8. "Universal" build resolves this compromise by providing
5926 near-optimal performance even on newer platforms.
5930 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5931 (other platforms pending).
5933 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5935 * Add support for the SignedCertificateTimestampList certificate and
5936 OCSP response extensions from RFC6962.
5940 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5941 for corner cases. (Certain input points at infinity could lead to
5942 bogus results, with non-infinity inputs mapped to infinity too.)
5946 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5947 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5948 common cases are optimized and there still is room for further
5949 improvements. Vector Permutation AES for Altivec is also added.
5953 * Add support for little-endian ppc64 Linux target.
5955 *Marcelo Cerri (IBM)*
5957 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5958 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5959 are optimized and there still is room for further improvements.
5960 Both 32- and 64-bit modes are supported.
5962 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5964 * Improved ARMv7 NEON support.
5968 * Support for SPARC Architecture 2011 crypto extensions, first
5969 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5970 SHA256/512, MD5, GHASH and modular exponentiation.
5972 *Andy Polyakov, David Miller*
5974 * Accelerated modular exponentiation for Intel processors, a.k.a.
5977 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5979 * Support for new and upcoming Intel processors, including AVX2,
5980 BMI and SHA ISA extensions. This includes additional "stitched"
5981 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5984 This work was sponsored by Intel Corp.
5988 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5989 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5990 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5994 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5995 this fixes a limitation in previous versions of OpenSSL.
5999 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6000 MGF1 digest and OAEP label.
6004 * Add EVP support for key wrapping algorithms, to avoid problems with
6005 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6006 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6007 algorithms and include tests cases.
6011 * Add functions to allocate and set the fields of an ECDSA_METHOD
6014 *Douglas E. Engert, Steve Henson*
6016 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6017 difference in days and seconds between two tm or ASN1_TIME structures.
6021 * Add -rev test option to s_server to just reverse order of characters
6022 received by client and send back to server. Also prints an abbreviated
6023 summary of the connection parameters.
6027 * New option -brief for s_client and s_server to print out a brief summary
6028 of connection parameters.
6032 * Add callbacks for arbitrary TLS extensions.
6034 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6036 * New option -crl_download in several openssl utilities to download CRLs
6037 from CRLDP extension in certificates.
6041 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6045 * New function X509_CRL_diff to generate a delta CRL from the difference
6046 of two full CRLs. Add support to "crl" utility.
6050 * New functions to set lookup_crls function and to retrieve
6051 X509_STORE from X509_STORE_CTX.
6055 * Print out deprecated issuer and subject unique ID fields in
6060 * Extend OCSP I/O functions so they can be used for simple general purpose
6061 HTTP as well as OCSP. New wrapper function which can be used to download
6062 CRLs using the OCSP API.
6066 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6070 * `SSL_CONF*` functions. These provide a common framework for application
6071 configuration using configuration files or command lines.
6075 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6076 message callback and prints the results. Needs compile time option
6077 "enable-ssl-trace". New options to s_client and s_server to enable
6082 * New ctrl and macro to retrieve supported points extensions.
6083 Print out extension in s_server and s_client.
6087 * New functions to retrieve certificate signature and signature
6092 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6097 * New Suite B modes for TLS code. These use and enforce the requirements
6098 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6099 only use Suite B curves. The Suite B modes can be set by using the
6100 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6104 * New chain verification flags for Suite B levels of security. Check
6105 algorithms are acceptable when flags are set in X509_verify_cert.
6109 * Make tls1_check_chain return a set of flags indicating checks passed
6110 by a certificate chain. Add additional tests to handle client
6111 certificates: checks for matching certificate type and issuer name
6116 * If an attempt is made to use a signature algorithm not in the peer
6117 preference list abort the handshake. If client has no suitable
6118 signature algorithms in response to a certificate request do not
6119 use the certificate.
6123 * If server EC tmp key is not in client preference list abort handshake.
6127 * Add support for certificate stores in CERT structure. This makes it
6128 possible to have different stores per SSL structure or one store in
6129 the parent SSL_CTX. Include distinct stores for certificate chain
6130 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6131 to build and store a certificate chain in CERT structure: returning
6132 an error if the chain cannot be built: this will allow applications
6133 to test if a chain is correctly configured.
6135 Note: if the CERT based stores are not set then the parent SSL_CTX
6136 store is used to retain compatibility with existing behaviour.
6140 * New function ssl_set_client_disabled to set a ciphersuite disabled
6141 mask based on the current session, check mask when sending client
6142 hello and checking the requested ciphersuite.
6146 * New ctrls to retrieve and set certificate types in a certificate
6147 request message. Print out received values in s_client. If certificate
6148 types is not set with custom values set sensible values based on
6149 supported signature algorithms.
6153 * Support for distinct client and server supported signature algorithms.
6157 * Add certificate callback. If set this is called whenever a certificate
6158 is required by client or server. An application can decide which
6159 certificate chain to present based on arbitrary criteria: for example
6160 supported signature algorithms. Add very simple example to s_server.
6161 This fixes many of the problems and restrictions of the existing client
6162 certificate callback: for example you can now clear an existing
6163 certificate and specify the whole chain.
6167 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6168 the certificate can be used for (if anything). Set valid_flags field
6169 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6170 to have similar checks in it.
6172 Add new "cert_flags" field to CERT structure and include a "strict mode".
6173 This enforces some TLS certificate requirements (such as only permitting
6174 certificate signature algorithms contained in the supported algorithms
6175 extension) which some implementations ignore: this option should be used
6176 with caution as it could cause interoperability issues.
6180 * Update and tidy signature algorithm extension processing. Work out
6181 shared signature algorithms based on preferences and peer algorithms
6182 and print them out in s_client and s_server. Abort handshake if no
6183 shared signature algorithms.
6187 * Add new functions to allow customised supported signature algorithms
6188 for SSL and SSL_CTX structures. Add options to s_client and s_server
6193 * New function SSL_certs_clear() to delete all references to certificates
6194 from an SSL structure. Before this once a certificate had been added
6195 it couldn't be removed.
6199 * Integrate hostname, email address and IP address checking with certificate
6200 verification. New verify options supporting checking in openssl utility.
6204 * Fixes and wildcard matching support to hostname and email checking
6205 functions. Add manual page.
6207 *Florian Weimer (Red Hat Product Security Team)*
6209 * New functions to check a hostname email or IP address against a
6210 certificate. Add options x509 utility to print results of checks against
6215 * Fix OCSP checking.
6217 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6219 * Initial experimental support for explicitly trusted non-root CAs.
6220 OpenSSL still tries to build a complete chain to a root but if an
6221 intermediate CA has a trust setting included that is used. The first
6222 setting is used: whether to trust (e.g., -addtrust option to the x509
6227 * Add -trusted_first option which attempts to find certificates in the
6228 trusted store even if an untrusted chain is also supplied.
6232 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6233 platform support for Linux and Android.
6237 * Support for linux-x32, ILP32 environment in x86_64 framework.
6241 * Experimental multi-implementation support for FIPS capable OpenSSL.
6242 When in FIPS mode the approved implementations are used as normal,
6243 when not in FIPS mode the internal unapproved versions are used instead.
6244 This means that the FIPS capable OpenSSL isn't forced to use the
6245 (often lower performance) FIPS implementations outside FIPS mode.
6249 * Transparently support X9.42 DH parameters when calling
6250 PEM_read_bio_DHparameters. This means existing applications can handle
6251 the new parameter format automatically.
6255 * Initial experimental support for X9.42 DH parameter format: mainly
6256 to support use of 'q' parameter for RFC5114 parameters.
6260 * Add DH parameters from RFC5114 including test data to dhtest.
6264 * Support for automatic EC temporary key parameter selection. If enabled
6265 the most preferred EC parameters are automatically used instead of
6266 hardcoded fixed parameters. Now a server just has to call:
6267 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6268 support ECDH and use the most appropriate parameters.
6272 * Enhance and tidy EC curve and point format TLS extension code. Use
6273 static structures instead of allocation if default values are used.
6274 New ctrls to set curves we wish to support and to retrieve shared curves.
6275 Print out shared curves in s_server. New options to s_server and s_client
6276 to set list of supported curves.
6280 * New ctrls to retrieve supported signature algorithms and
6281 supported curve values as an array of NIDs. Extend openssl utility
6282 to print out received values.
6286 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6287 between NIDs and the more common NIST names such as "P-256". Enhance
6288 ecparam utility and ECC method to recognise the NIST names for curves.
6292 * Enhance SSL/TLS certificate chain handling to support different
6293 chains for each certificate instead of one chain in the parent SSL_CTX.
6297 * Support for fixed DH ciphersuite client authentication: where both
6298 server and client use DH certificates with common parameters.
6302 * Support for fixed DH ciphersuites: those requiring DH server
6307 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6309 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6310 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6311 X509_CINF_get_signature were reverted post internal team review.
6316 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6318 * OCSP Status Request extension unbounded memory growth
6320 A malicious client can send an excessively large OCSP Status Request
6321 extension. If that client continually requests renegotiation, sending a
6322 large OCSP Status Request extension each time, then there will be unbounded
6323 memory growth on the server. This will eventually lead to a Denial Of
6324 Service attack through memory exhaustion. Servers with a default
6325 configuration are vulnerable even if they do not support OCSP. Builds using
6326 the "no-ocsp" build time option are not affected.
6328 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6333 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6336 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6342 * OOB write in MDC2_Update()
6344 An overflow can occur in MDC2_Update() either if called directly or
6345 through the EVP_DigestUpdate() function using MDC2. If an attacker
6346 is able to supply very large amounts of input data after a previous
6347 call to EVP_EncryptUpdate() with a partial block then a length check
6348 can overflow resulting in a heap corruption.
6350 The amount of data needed is comparable to SIZE_MAX which is impractical
6353 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6358 * Malformed SHA512 ticket DoS
6360 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6361 DoS attack where a malformed ticket will result in an OOB read which will
6364 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6365 a custom server callback and ticket lookup mechanism.
6367 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6372 * OOB write in BN_bn2dec()
6374 The function BN_bn2dec() does not check the return value of BN_div_word().
6375 This can cause an OOB write if an application uses this function with an
6376 overly large BIGNUM. This could be a problem if an overly large certificate
6377 or CRL is printed out from an untrusted source. TLS is not affected because
6378 record limits will reject an oversized certificate before it is parsed.
6380 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6385 * OOB read in TS_OBJ_print_bio()
6387 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6388 the total length the OID text representation would use and not the amount
6389 of data written. This will result in OOB reads when large OIDs are
6392 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6397 * Pointer arithmetic undefined behaviour
6399 Avoid some undefined pointer arithmetic
6401 A common idiom in the codebase is to check limits in the following manner:
6404 Where "p" points to some malloc'd data of SIZE bytes and
6407 "len" here could be from some externally supplied data (e.g. from a TLS
6410 The rules of C pointer arithmetic are such that "p + len" is only well
6411 defined where len <= SIZE. Therefore the above idiom is actually
6412 undefined behaviour.
6414 For example this could cause problems if some malloc implementation
6415 provides an address for "p" such that "p + len" actually overflows for
6416 values of len that are too big and therefore p + len < limit.
6418 This issue was reported to OpenSSL by Guido Vranken
6423 * Constant time flag not preserved in DSA signing
6425 Operations in the DSA signing algorithm should run in constant time in
6426 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6427 implementation means that a non-constant time codepath is followed for
6428 certain operations. This has been demonstrated through a cache-timing
6429 attack to be sufficient for an attacker to recover the private DSA key.
6431 This issue was reported by César Pereida (Aalto University), Billy Brumley
6432 (Tampere University of Technology), and Yuval Yarom (The University of
6433 Adelaide and NICTA).
6438 * DTLS buffered message DoS
6440 In a DTLS connection where handshake messages are delivered out-of-order
6441 those messages that OpenSSL is not yet ready to process will be buffered
6442 for later use. Under certain circumstances, a flaw in the logic means that
6443 those messages do not get removed from the buffer even though the handshake
6444 has been completed. An attacker could force up to approx. 15 messages to
6445 remain in the buffer when they are no longer required. These messages will
6446 be cleared when the DTLS connection is closed. The default maximum size for
6447 a message is 100k. Therefore the attacker could force an additional 1500k
6448 to be consumed per connection. By opening many simulataneous connections an
6449 attacker could cause a DoS attack through memory exhaustion.
6451 This issue was reported to OpenSSL by Quan Luo.
6456 * DTLS replay protection DoS
6458 A flaw in the DTLS replay attack protection mechanism means that records
6459 that arrive for future epochs update the replay protection "window" before
6460 the MAC for the record has been validated. This could be exploited by an
6461 attacker by sending a record for the next epoch (which does not have to
6462 decrypt or have a valid MAC), with a very large sequence number. This means
6463 that all subsequent legitimate packets are dropped causing a denial of
6464 service for a specific DTLS connection.
6466 This issue was reported to OpenSSL by the OCAP audit team.
6471 * Certificate message OOB reads
6473 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6474 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6475 theoretical DoS risk but this has not been observed in practice on common
6478 The messages affected are client certificate, client certificate request
6479 and server certificate. As a result the attack can only be performed
6480 against a client or a server which enables client authentication.
6482 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6487 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6489 * Prevent padding oracle in AES-NI CBC MAC check
6491 A MITM attacker can use a padding oracle attack to decrypt traffic
6492 when the connection uses an AES CBC cipher and the server support
6495 This issue was introduced as part of the fix for Lucky 13 padding
6496 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6497 constant time by making sure that always the same bytes are read and
6498 compared against either the MAC or padding bytes. But it no longer
6499 checked that there was enough data to have both the MAC and padding
6502 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6507 * Fix EVP_EncodeUpdate overflow
6509 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6510 Base64 encoding of binary data. If an attacker is able to supply very large
6511 amounts of input data then a length check can overflow resulting in a heap
6514 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6515 the `PEM_write_bio*` family of functions. These are mainly used within the
6516 OpenSSL command line applications, so any application which processes data
6517 from an untrusted source and outputs it as a PEM file should be considered
6518 vulnerable to this issue. User applications that call these APIs directly
6519 with large amounts of untrusted data may also be vulnerable.
6521 This issue was reported by Guido Vranken.
6526 * Fix EVP_EncryptUpdate overflow
6528 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6529 is able to supply very large amounts of input data after a previous call to
6530 EVP_EncryptUpdate() with a partial block then a length check can overflow
6531 resulting in a heap corruption. Following an analysis of all OpenSSL
6532 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6533 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6534 the first called function after an EVP_EncryptInit(), and therefore that
6535 specific call must be safe. The second form is where the length passed to
6536 EVP_EncryptUpdate() can be seen from the code to be some small value and
6537 therefore there is no possibility of an overflow. Since all instances are
6538 one of these two forms, it is believed that there can be no overflows in
6539 internal code due to this problem. It should be noted that
6540 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6541 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6542 of these calls have also been analysed too and it is believed there are no
6543 instances in internal usage where an overflow could occur.
6545 This issue was reported by Guido Vranken.
6550 * Prevent ASN.1 BIO excessive memory allocation
6552 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6553 a short invalid encoding can casuse allocation of large amounts of memory
6554 potentially consuming excessive resources or exhausting memory.
6556 Any application parsing untrusted data through d2i BIO functions is
6557 affected. The memory based functions such as d2i_X509() are *not* affected.
6558 Since the memory based functions are used by the TLS library, TLS
6559 applications are not affected.
6561 This issue was reported by Brian Carpenter.
6568 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6569 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6570 in arbitrary stack data being returned in the buffer.
6572 This issue was reported by Guido Vranken.
6577 * Modify behavior of ALPN to invoke callback after SNI/servername
6578 callback, such that updates to the SSL_CTX affect ALPN.
6582 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6587 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6588 methods are enabled and ssl2 is disabled the methods return NULL.
6592 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6594 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6595 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6596 provide any "EXPORT" or "LOW" strength ciphers.
6600 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6601 is by default disabled at build-time. Builds that are not configured with
6602 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6603 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6604 will need to explicitly call either of:
6606 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6608 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6610 as appropriate. Even if either of those is used, or the application
6611 explicitly uses the version-specific SSLv2_method() or its client and
6612 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6613 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6614 ciphers, and SSLv2 56-bit DES are no longer available.
6619 * Fix a double-free in DSA code
6621 A double free bug was discovered when OpenSSL parses malformed DSA private
6622 keys and could lead to a DoS attack or memory corruption for applications
6623 that receive DSA private keys from untrusted sources. This scenario is
6626 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6632 * Disable SRP fake user seed to address a server memory leak.
6634 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6636 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6637 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6638 was changed to ignore the "fake user" SRP seed, even if the seed
6641 Users should use SRP_VBASE_get1_by_user instead. Note that in
6642 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6643 also that even though configuring the SRP seed attempts to hide
6644 invalid usernames by continuing the handshake with fake
6645 credentials, this behaviour is not constant time and no strong
6646 guarantees are made that the handshake is indistinguishable from
6647 that of a valid user.
6652 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6654 In the BN_hex2bn function the number of hex digits is calculated using an
6655 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6656 large values of `i` this can result in `bn_expand` not allocating any
6657 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6658 field as NULL leading to a subsequent NULL ptr deref. For very large values
6659 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6660 In this case memory is allocated to the internal BIGNUM data field, but it
6661 is insufficiently sized leading to heap corruption. A similar issue exists
6662 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6663 is ever called by user applications with very large untrusted hex/dec data.
6664 This is anticipated to be a rare occurrence.
6666 All OpenSSL internal usage of these functions use data that is not expected
6667 to be untrusted, e.g. config file data or application command line
6668 arguments. If user developed applications generate config file data based
6669 on untrusted data then it is possible that this could also lead to security
6670 consequences. This is also anticipated to be rare.
6672 This issue was reported to OpenSSL by Guido Vranken.
6677 * Fix memory issues in `BIO_*printf` functions
6679 The internal `fmtstr` function used in processing a "%s" format string in
6680 the `BIO_*printf` functions could overflow while calculating the length of a
6681 string and cause an OOB read when printing very long strings.
6683 Additionally the internal `doapr_outch` function can attempt to write to an
6684 OOB memory location (at an offset from the NULL pointer) in the event of a
6685 memory allocation failure. In 1.0.2 and below this could be caused where
6686 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6687 could be in processing a very long "%s" format string. Memory leaks can
6690 The first issue may mask the second issue dependent on compiler behaviour.
6691 These problems could enable attacks where large amounts of untrusted data
6692 is passed to the `BIO_*printf` functions. If applications use these functions
6693 in this way then they could be vulnerable. OpenSSL itself uses these
6694 functions when printing out human-readable dumps of ASN.1 data. Therefore
6695 applications that print this data could be vulnerable if the data is from
6696 untrusted sources. OpenSSL command line applications could also be
6697 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6698 as command line arguments.
6700 Libssl is not considered directly vulnerable. Additionally certificates etc
6701 received via remote connections via libssl are also unlikely to be able to
6702 trigger these issues because of message size limits enforced within libssl.
6704 This issue was reported to OpenSSL Guido Vranken.
6709 * Side channel attack on modular exponentiation
6711 A side-channel attack was found which makes use of cache-bank conflicts on
6712 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6713 of RSA keys. The ability to exploit this issue is limited as it relies on
6714 an attacker who has control of code in a thread running on the same
6715 hyper-threaded core as the victim thread which is performing decryptions.
6717 This issue was reported to OpenSSL by Yuval Yarom, The University of
6718 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6719 Nadia Heninger, University of Pennsylvania with more information at
6720 <http://cachebleed.info>.
6725 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6726 if no keysize is specified with default_bits. This fixes an
6727 omission in an earlier change that changed all RSA/DSA key generation
6728 commands to use 2048 bits by default.
6732 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6734 * Protection for DH small subgroup attacks
6736 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6737 switched on by default and cannot be disabled. This could have some
6742 * SSLv2 doesn't block disabled ciphers
6744 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6745 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6746 been disabled, provided that the SSLv2 protocol was not also disabled via
6749 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6750 and Sebastian Schinzel.
6755 * Reject DH handshakes with parameters shorter than 1024 bits.
6759 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6761 * Certificate verify crash with missing PSS parameter
6763 The signature verification routines will crash with a NULL pointer
6764 dereference if presented with an ASN.1 signature using the RSA PSS
6765 algorithm and absent mask generation function parameter. Since these
6766 routines are used to verify certificate signature algorithms this can be
6767 used to crash any certificate verification operation and exploited in a
6768 DoS attack. Any application which performs certificate verification is
6769 vulnerable including OpenSSL clients and servers which enable client
6772 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6777 * X509_ATTRIBUTE memory leak
6779 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6780 memory. This structure is used by the PKCS#7 and CMS routines so any
6781 application which reads PKCS#7 or CMS data from untrusted sources is
6782 affected. SSL/TLS is not affected.
6784 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6790 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6791 This changes the decoding behaviour for some invalid messages,
6792 though the change is mostly in the more lenient direction, and
6793 legacy behaviour is preserved as much as possible.
6797 * In DSA_generate_parameters_ex, if the provided seed is too short,
6798 use a random seed, as already documented.
6800 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6802 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6804 * Alternate chains certificate forgery
6806 During certificate verfification, OpenSSL will attempt to find an
6807 alternative certificate chain if the first attempt to build such a chain
6808 fails. An error in the implementation of this logic can mean that an
6809 attacker could cause certain checks on untrusted certificates to be
6810 bypassed, such as the CA flag, enabling them to use a valid leaf
6811 certificate to act as a CA and "issue" an invalid certificate.
6813 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6819 * Race condition handling PSK identify hint
6821 If PSK identity hints are received by a multi-threaded client then
6822 the values are wrongly updated in the parent SSL_CTX structure. This can
6823 result in a race condition potentially leading to a double free of the
6829 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6831 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6832 incompatibility in the handling of HMAC. The previous ABI has now been
6835 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6837 * Malformed ECParameters causes infinite loop
6839 When processing an ECParameters structure OpenSSL enters an infinite loop
6840 if the curve specified is over a specially malformed binary polynomial
6843 This can be used to perform denial of service against any
6844 system which processes public keys, certificate requests or
6845 certificates. This includes TLS clients and TLS servers with
6846 client authentication enabled.
6848 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6853 * Exploitable out-of-bounds read in X509_cmp_time
6855 X509_cmp_time does not properly check the length of the ASN1_TIME
6856 string and can read a few bytes out of bounds. In addition,
6857 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6860 An attacker can use this to craft malformed certificates and CRLs of
6861 various sizes and potentially cause a segmentation fault, resulting in
6862 a DoS on applications that verify certificates or CRLs. TLS clients
6863 that verify CRLs are affected. TLS clients and servers with client
6864 authentication enabled may be affected if they use custom verification
6867 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6868 independently by Hanno Böck.
6873 * PKCS7 crash with missing EnvelopedContent
6875 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6876 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6877 with missing content and trigger a NULL pointer dereference on parsing.
6879 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6880 structures from untrusted sources are affected. OpenSSL clients and
6881 servers are not affected.
6883 This issue was reported to OpenSSL by Michal Zalewski (Google).
6888 * CMS verify infinite loop with unknown hash function
6890 When verifying a signedData message the CMS code can enter an infinite loop
6891 if presented with an unknown hash function OID. This can be used to perform
6892 denial of service against any system which verifies signedData messages using
6894 This issue was reported to OpenSSL by Johannes Bauer.
6899 * Race condition handling NewSessionTicket
6901 If a NewSessionTicket is received by a multi-threaded client when attempting to
6902 reuse a previous ticket then a race condition can occur potentially leading to
6903 a double free of the ticket data.
6908 * Reject DH handshakes with parameters shorter than 768 bits.
6910 *Kurt Roeckx and Emilia Kasper*
6912 * dhparam: generate 2048-bit parameters by default.
6914 *Kurt Roeckx and Emilia Kasper*
6916 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6918 * Segmentation fault in ASN1_TYPE_cmp fix
6920 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6921 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6922 certificate signature algorithm consistency this can be used to crash any
6923 certificate verification operation and exploited in a DoS attack. Any
6924 application which performs certificate verification is vulnerable including
6925 OpenSSL clients and servers which enable client authentication.
6930 * ASN.1 structure reuse memory corruption fix
6932 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6933 memory corruption via an invalid write. Such reuse is and has been
6934 strongly discouraged and is believed to be rare.
6936 Applications that parse structures containing CHOICE or ANY DEFINED BY
6937 components may be affected. Certificate parsing (d2i_X509 and related
6938 functions) are however not affected. OpenSSL clients and servers are
6944 * PKCS7 NULL pointer dereferences fix
6946 The PKCS#7 parsing code does not handle missing outer ContentInfo
6947 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6948 missing content and trigger a NULL pointer dereference on parsing.
6950 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6951 otherwise parse PKCS#7 structures from untrusted sources are
6952 affected. OpenSSL clients and servers are not affected.
6954 This issue was reported to OpenSSL by Michal Zalewski (Google).
6959 * DoS via reachable assert in SSLv2 servers fix
6961 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6962 servers that both support SSLv2 and enable export cipher suites by sending
6963 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6965 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6966 (OpenSSL development team).
6971 * Use After Free following d2i_ECPrivatekey error fix
6973 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6974 could cause a use after free condition. This, in turn, could cause a double
6975 free in several private key parsing functions (such as d2i_PrivateKey
6976 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6977 for applications that receive EC private keys from untrusted
6978 sources. This scenario is considered rare.
6980 This issue was discovered by the BoringSSL project and fixed in their
6986 * X509_to_X509_REQ NULL pointer deref fix
6988 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6989 the certificate key is invalid. This function is rarely used in practice.
6991 This issue was discovered by Brian Carpenter.
6996 * Removed the export ciphers from the DEFAULT ciphers
7000 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7002 * Build fixes for the Windows and OpenVMS platforms
7004 *Matt Caswell and Richard Levitte*
7006 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7008 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7009 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7010 dereference. This could lead to a Denial Of Service attack. Thanks to
7011 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7016 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7017 dtls1_buffer_record function under certain conditions. In particular this
7018 could occur if an attacker sent repeated DTLS records with the same
7019 sequence number but for the next epoch. The memory leak could be exploited
7020 by an attacker in a Denial of Service attack through memory exhaustion.
7021 Thanks to Chris Mueller for reporting this issue.
7026 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7027 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7028 method would be set to NULL which could later result in a NULL pointer
7029 dereference. Thanks to Frank Schmirler for reporting this issue.
7034 * Abort handshake if server key exchange message is omitted for ephemeral
7037 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7038 reporting this issue.
7043 * Remove non-export ephemeral RSA code on client and server. This code
7044 violated the TLS standard by allowing the use of temporary RSA keys in
7045 non-export ciphersuites and could be used by a server to effectively
7046 downgrade the RSA key length used to a value smaller than the server
7047 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7048 INRIA or reporting this issue.
7053 * Fixed issue where DH client certificates are accepted without verification.
7054 An OpenSSL server will accept a DH certificate for client authentication
7055 without the certificate verify message. This effectively allows a client to
7056 authenticate without the use of a private key. This only affects servers
7057 which trust a client certificate authority which issues certificates
7058 containing DH keys: these are extremely rare and hardly ever encountered.
7059 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7065 * Ensure that the session ID context of an SSL is updated when its
7066 SSL_CTX is updated via SSL_set_SSL_CTX.
7068 The session ID context is typically set from the parent SSL_CTX,
7069 and can vary with the CTX.
7073 * Fix various certificate fingerprint issues.
7075 By using non-DER or invalid encodings outside the signed portion of a
7076 certificate the fingerprint can be changed without breaking the signature.
7077 Although no details of the signed portion of the certificate can be changed
7078 this can cause problems with some applications: e.g. those using the
7079 certificate fingerprint for blacklists.
7081 1. Reject signatures with non zero unused bits.
7083 If the BIT STRING containing the signature has non zero unused bits reject
7084 the signature. All current signature algorithms require zero unused bits.
7086 2. Check certificate algorithm consistency.
7088 Check the AlgorithmIdentifier inside TBS matches the one in the
7089 certificate signature. NB: this will result in signature failure
7090 errors for some broken certificates.
7092 Thanks to Konrad Kraszewski from Google for reporting this issue.
7094 3. Check DSA/ECDSA signatures use DER.
7096 Re-encode DSA/ECDSA signatures and compare with the original received
7097 signature. Return an error if there is a mismatch.
7099 This will reject various cases including garbage after signature
7100 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7101 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7102 (negative or with leading zeroes).
7104 Further analysis was conducted and fixes were developed by Stephen Henson
7105 of the OpenSSL core team.
7111 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7112 results on some platforms, including x86_64. This bug occurs at random
7113 with a very low probability, and is not known to be exploitable in any
7114 way, though its exact impact is difficult to determine. Thanks to Pieter
7115 Wuille (Blockstream) who reported this issue and also suggested an initial
7116 fix. Further analysis was conducted by the OpenSSL development team and
7117 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7118 the OpenSSL core team.
7123 * Do not resume sessions on the server if the negotiated protocol
7124 version does not match the session's version. Resuming with a different
7125 version, while not strictly forbidden by the RFC, is of questionable
7126 sanity and breaks all known clients.
7128 *David Benjamin, Emilia Käsper*
7130 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7131 early CCS messages during renegotiation. (Note that because
7132 renegotiation is encrypted, this early CCS was not exploitable.)
7136 * Tighten client-side session ticket handling during renegotiation:
7137 ensure that the client only accepts a session ticket if the server sends
7138 the extension anew in the ServerHello. Previously, a TLS client would
7139 reuse the old extension state and thus accept a session ticket if one was
7140 announced in the initial ServerHello.
7142 Similarly, ensure that the client requires a session ticket if one
7143 was advertised in the ServerHello. Previously, a TLS client would
7144 ignore a missing NewSessionTicket message.
7148 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7152 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7153 sends a carefully crafted handshake message, to cause OpenSSL to fail
7154 to free up to 64k of memory causing a memory leak. This could be
7155 exploited in a Denial Of Service attack. This issue affects OpenSSL
7156 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7157 whether SRTP is used or configured. Implementations of OpenSSL that
7158 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7160 The fix was developed by the OpenSSL team.
7165 * Session Ticket Memory Leak.
7167 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7168 integrity of that ticket is first verified. In the event of a session
7169 ticket integrity check failing, OpenSSL will fail to free memory
7170 causing a memory leak. By sending a large number of invalid session
7171 tickets an attacker could exploit this issue in a Denial Of Service
7177 * Build option no-ssl3 is incomplete.
7179 When OpenSSL is configured with "no-ssl3" as a build option, servers
7180 could accept and complete a SSL 3.0 handshake, and clients could be
7181 configured to send them.
7184 *Akamai and the OpenSSL team*
7186 * Add support for TLS_FALLBACK_SCSV.
7187 Client applications doing fallback retries should call
7188 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7191 *Adam Langley, Bodo Moeller*
7193 * Add additional DigestInfo checks.
7195 Re-encode DigestInto in DER and check against the original when
7196 verifying RSA signature: this will reject any improperly encoded
7197 DigestInfo structures.
7199 Note: this is a precautionary measure and no attacks are currently known.
7203 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7205 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7206 SRP code can be overrun an internal buffer. Add sanity check that
7207 g, A, B < N to SRP code.
7209 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7210 Group for discovering this issue.
7215 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7216 TLS 1.0 instead of higher protocol versions when the ClientHello message
7217 is badly fragmented. This allows a man-in-the-middle attacker to force a
7218 downgrade to TLS 1.0 even if both the server and the client support a
7219 higher protocol version, by modifying the client's TLS records.
7221 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7222 researching this issue.
7227 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7228 to a denial of service attack. A malicious server can crash the client
7229 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7230 ciphersuite and sending carefully crafted handshake messages.
7232 Thanks to Felix Gröbert (Google) for discovering and researching this
7238 * By sending carefully crafted DTLS packets an attacker could cause openssl
7239 to leak memory. This can be exploited through a Denial of Service attack.
7240 Thanks to Adam Langley for discovering and researching this issue.
7245 * An attacker can force openssl to consume large amounts of memory whilst
7246 processing DTLS handshake messages. This can be exploited through a
7247 Denial of Service attack.
7248 Thanks to Adam Langley for discovering and researching this issue.
7253 * An attacker can force an error condition which causes openssl to crash
7254 whilst processing DTLS packets due to memory being freed twice. This
7255 can be exploited through a Denial of Service attack.
7256 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7262 * If a multithreaded client connects to a malicious server using a resumed
7263 session and the server sends an ec point format extension it could write
7264 up to 255 bytes to freed memory.
7266 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7272 * A malicious server can crash an OpenSSL client with a null pointer
7273 dereference (read) by specifying an SRP ciphersuite even though it was not
7274 properly negotiated with the client. This can be exploited through a
7275 Denial of Service attack.
7277 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7278 discovering and researching this issue.
7283 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7284 X509_name_oneline, X509_name_print_ex et al. to leak some information
7285 from the stack. Applications may be affected if they echo pretty printing
7286 output to the attacker.
7288 Thanks to Ivan Fratric (Google) for discovering this issue.
7291 *Emilia Käsper, and Steve Henson*
7293 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7294 for corner cases. (Certain input points at infinity could lead to
7295 bogus results, with non-infinity inputs mapped to infinity too.)
7299 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7301 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7302 handshake can force the use of weak keying material in OpenSSL
7303 SSL/TLS clients and servers.
7305 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7306 researching this issue. ([CVE-2014-0224])
7308 *KIKUCHI Masashi, Steve Henson*
7310 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7311 OpenSSL DTLS client the code can be made to recurse eventually crashing
7314 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7317 *Imre Rad, Steve Henson*
7319 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7320 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7321 client or server. This is potentially exploitable to run arbitrary
7322 code on a vulnerable client or server.
7324 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7326 *Jüri Aedla, Steve Henson*
7328 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7329 are subject to a denial of service attack.
7331 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7332 this issue. ([CVE-2014-3470])
7334 *Felix Gröbert, Ivan Fratric, Steve Henson*
7336 * Harmonize version and its documentation. -f flag is used to display
7339 *mancha <mancha1@zoho.com>*
7341 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7342 in i2d_ECPrivateKey.
7344 *mancha <mancha1@zoho.com>*
7346 * Fix some double frees. These are not thought to be exploitable.
7348 *mancha <mancha1@zoho.com>*
7350 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7352 * A missing bounds check in the handling of the TLS heartbeat extension
7353 can be used to reveal up to 64k of memory to a connected client or
7356 Thanks for Neel Mehta of Google Security for discovering this bug and to
7357 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7358 preparing the fix ([CVE-2014-0160])
7360 *Adam Langley, Bodo Moeller*
7362 * Fix for the attack described in the paper "Recovering OpenSSL
7363 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7364 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7365 <http://eprint.iacr.org/2014/140>
7367 Thanks to Yuval Yarom and Naomi Benger for discovering this
7368 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7370 *Yuval Yarom and Naomi Benger*
7372 * TLS pad extension: draft-agl-tls-padding-03
7374 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7375 TLS client Hello record length value would otherwise be > 255 and
7376 less that 512 pad with a dummy extension containing zeroes so it
7377 is at least 512 bytes long.
7379 *Adam Langley, Steve Henson*
7381 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7383 * Fix for TLS record tampering bug. A carefully crafted invalid
7384 handshake could crash OpenSSL with a NULL pointer exception.
7385 Thanks to Anton Johansson for reporting this issues.
7388 * Keep original DTLS digest and encryption contexts in retransmission
7389 structures so we can use the previous session parameters if they need
7390 to be resent. ([CVE-2013-6450])
7394 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7395 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7396 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7397 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7398 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7399 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7401 *Rob Stradling, Adam Langley*
7403 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7405 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7406 supporting platforms or when small records were transferred.
7408 *Andy Polyakov, Steve Henson*
7410 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7412 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7414 This addresses the flaw in CBC record processing discovered by
7415 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7416 at: <http://www.isg.rhul.ac.uk/tls/>
7418 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7419 Security Group at Royal Holloway, University of London
7420 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7421 Emilia Käsper for the initial patch.
7424 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7426 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7427 ciphersuites which can be exploited in a denial of service attack.
7428 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7429 and detecting this bug and to Wolfgang Ettlinger
7430 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7435 * Return an error when checking OCSP signatures when key is NULL.
7436 This fixes a DoS attack. ([CVE-2013-0166])
7440 * Make openssl verify return errors.
7442 *Chris Palmer <palmer@google.com> and Ben Laurie*
7444 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7445 the right response is stapled. Also change SSL_get_certificate()
7446 so it returns the certificate actually sent.
7447 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7449 *Rob Stradling <rob.stradling@comodo.com>*
7451 * Fix possible deadlock when decoding public keys.
7455 * Don't use TLS 1.0 record version number in initial client hello
7460 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7462 * Sanity check record length before skipping explicit IV in TLS
7463 1.2, 1.1 and DTLS to fix DoS attack.
7465 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7466 fuzzing as a service testing platform.
7471 * Initialise tkeylen properly when encrypting CMS messages.
7472 Thanks to Solar Designer of Openwall for reporting this issue.
7476 * In FIPS mode don't try to use composite ciphers as they are not
7481 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7483 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7484 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7485 mean any application compiled against OpenSSL 1.0.0 headers setting
7486 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7487 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7488 0x10000000L Any application which was previously compiled against
7489 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7490 will need to be recompiled as a result. Letting be results in
7491 inability to disable specifically TLS 1.1 and in client context,
7492 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7496 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7497 disable just protocol X, but all protocols above X *if* there are
7498 protocols *below* X still enabled. In more practical terms it means
7499 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7500 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7501 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7506 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7508 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7509 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7510 in CRYPTO_realloc_clean.
7512 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7513 issue and to Adam Langley <agl@chromium.org> for fixing it.
7516 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7518 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7522 * Workarounds for some broken servers that "hang" if a client hello
7523 record length exceeds 255 bytes.
7525 1. Do not use record version number > TLS 1.0 in initial client
7526 hello: some (but not all) hanging servers will now work.
7527 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7528 the number of ciphers sent in the client hello. This should be
7529 set to an even number, such as 50, for example by passing:
7530 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7531 Most broken servers should now work.
7532 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7533 TLS 1.2 client support entirely.
7537 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7541 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7543 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7544 STRING form instead of a DigestInfo.
7548 * The format used for MDC2 RSA signatures is inconsistent between EVP
7549 and the RSA_sign/RSA_verify functions. This was made more apparent when
7550 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7551 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7552 the correct format in RSA_verify so both forms transparently work.
7556 * Some servers which support TLS 1.0 can choke if we initially indicate
7557 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7558 encrypted premaster secret. As a workaround use the maximum permitted
7559 client version in client hello, this should keep such servers happy
7560 and still work with previous versions of OpenSSL.
7564 * Add support for TLS/DTLS heartbeats.
7566 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7568 * Add support for SCTP.
7570 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7572 * Improved PRNG seeding for VOS.
7574 *Paul Green <Paul.Green@stratus.com>*
7576 * Extensive assembler packs updates, most notably:
7578 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7579 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7580 - x86_64: bit-sliced AES implementation;
7581 - ARM: NEON support, contemporary platforms optimizations;
7582 - s390x: z196 support;
7583 - `*`: GHASH and GF(2^m) multiplication implementations;
7587 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7588 (removal of unnecessary code)
7590 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7592 * Add TLS key material exporter from RFC 5705.
7596 * Add DTLS-SRTP negotiation from RFC 5764.
7600 * Add Next Protocol Negotiation,
7601 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7602 disabled with a no-npn flag to config or Configure. Code donated
7605 *Adam Langley <agl@google.com> and Ben Laurie*
7607 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7608 NIST-P256, NIST-P521, with constant-time single point multiplication on
7609 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7610 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7611 Code made available under Apache License version 2.0.
7613 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7614 line to include this in your build of OpenSSL, and run "make depend" (or
7615 "make update"). This enables the following EC_METHODs:
7617 EC_GFp_nistp224_method()
7618 EC_GFp_nistp256_method()
7619 EC_GFp_nistp521_method()
7621 EC_GROUP_new_by_curve_name() will automatically use these (while
7622 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7625 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7627 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7628 all platforms. Move ssize_t definition from e_os.h to the public
7629 header file e_os2.h as it now appears in public header file cms.h
7633 * New -sigopt option to the ca, req and x509 utilities. Additional
7634 signature parameters can be passed using this option and in
7639 * Add RSA PSS signing function. This will generate and set the
7640 appropriate AlgorithmIdentifiers for PSS based on those in the
7641 corresponding EVP_MD_CTX structure. No application support yet.
7645 * Support for companion algorithm specific ASN1 signing routines.
7646 New function ASN1_item_sign_ctx() signs a pre-initialised
7647 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7648 the appropriate parameters.
7652 * Add new algorithm specific ASN1 verification initialisation function
7653 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7654 handling will be the same no matter what EVP_PKEY_METHOD is used.
7655 Add a PSS handler to support verification of PSS signatures: checked
7656 against a number of sample certificates.
7660 * Add signature printing for PSS. Add PSS OIDs.
7662 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7664 * Add algorithm specific signature printing. An individual ASN1 method
7665 can now print out signatures instead of the standard hex dump.
7667 More complex signatures (e.g. PSS) can print out more meaningful
7668 information. Include DSA version that prints out the signature
7673 * Password based recipient info support for CMS library: implementing
7678 * Split password based encryption into PBES2 and PBKDF2 functions. This
7679 neatly separates the code into cipher and PBE sections and is required
7680 for some algorithms that split PBES2 into separate pieces (such as
7681 password based CMS).
7685 * Session-handling fixes:
7686 - Fix handling of connections that are resuming with a session ID,
7687 but also support Session Tickets.
7688 - Fix a bug that suppressed issuing of a new ticket if the client
7689 presented a ticket with an expired session.
7690 - Try to set the ticket lifetime hint to something reasonable.
7691 - Make tickets shorter by excluding irrelevant information.
7692 - On the client side, don't ignore renewed tickets.
7694 *Adam Langley, Bodo Moeller (Google)*
7696 * Fix PSK session representation.
7700 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7702 This work was sponsored by Intel.
7706 * Add GCM support to TLS library. Some custom code is needed to split
7707 the IV between the fixed (from PRF) and explicit (from TLS record)
7708 portions. This adds all GCM ciphersuites supported by RFC5288 and
7709 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7710 add a special AESGCM string for GCM only.
7714 * Expand range of ctrls for AES GCM. Permit setting invocation
7715 field on decrypt and retrieval of invocation field only on encrypt.
7719 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7720 As required by RFC5289 these ciphersuites cannot be used if for
7721 versions of TLS earlier than 1.2.
7725 * For FIPS capable OpenSSL interpret a NULL default public key method
7726 as unset and return the appropriate default but do *not* set the default.
7727 This means we can return the appropriate method in applications that
7728 switch between FIPS and non-FIPS modes.
7732 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7733 ENGINE is used then we cannot handle that in the FIPS module so we
7734 keep original code iff non-FIPS operations are allowed.
7738 * Add -attime option to openssl utilities.
7740 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7742 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7746 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7747 FIPS EC methods unconditionally for now.
7751 * New build option no-ec2m to disable characteristic 2 code.
7755 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7756 all cases can be covered as some introduce binary incompatibilities.
7760 * Redirect RSA operations to FIPS module including keygen,
7761 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7765 * Add similar low-level API blocking to ciphers.
7769 * low-level digest APIs are not approved in FIPS mode: any attempt
7770 to use these will cause a fatal error. Applications that *really* want
7771 to use them can use the `private_*` version instead.
7775 * Redirect cipher operations to FIPS module for FIPS builds.
7779 * Redirect digest operations to FIPS module for FIPS builds.
7783 * Update build system to add "fips" flag which will link in fipscanister.o
7784 for static and shared library builds embedding a signature if needed.
7788 * Output TLS supported curves in preference order instead of numerical
7789 order. This is currently hardcoded for the highest order curves first.
7790 This should be configurable so applications can judge speed vs strength.
7794 * Add TLS v1.2 server support for client authentication.
7798 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7803 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7804 FIPS modules versions.
7808 * Add TLS v1.2 client side support for client authentication. Keep cache
7809 of handshake records longer as we don't know the hash algorithm to use
7810 until after the certificate request message is received.
7814 * Initial TLS v1.2 client support. Add a default signature algorithms
7815 extension including all the algorithms we support. Parse new signature
7816 format in client key exchange. Relax some ECC signing restrictions for
7817 TLS v1.2 as indicated in RFC5246.
7821 * Add server support for TLS v1.2 signature algorithms extension. Switch
7822 to new signature format when needed using client digest preference.
7823 All server ciphersuites should now work correctly in TLS v1.2. No client
7824 support yet and no support for client certificates.
7828 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7829 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7830 ciphersuites. At present only RSA key exchange ciphersuites work with
7831 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7832 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7833 and version checking.
7837 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7838 with this defined it will not be affected by any changes to ssl internal
7839 structures. Add several utility functions to allow openssl application
7840 to work with OPENSSL_NO_SSL_INTERN defined.
7844 * A long standing patch to add support for SRP from EdelWeb (Peter
7845 Sylvester and Christophe Renou) was integrated.
7846 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7847 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7850 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7854 * Permit abbreviated handshakes when renegotiating using the function
7855 SSL_renegotiate_abbreviated().
7857 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7859 * Add call to ENGINE_register_all_complete() to
7860 ENGINE_load_builtin_engines(), so some implementations get used
7861 automatically instead of needing explicit application support.
7865 * Add support for TLS key exporter as described in RFC5705.
7867 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7869 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7870 a few changes are required:
7872 Add SSL_OP_NO_TLSv1_1 flag.
7873 Add TLSv1_1 methods.
7874 Update version checking logic to handle version 1.1.
7875 Add explicit IV handling (ported from DTLS code).
7876 Add command line options to s_client/s_server.
7883 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7885 * X509_ATTRIBUTE memory leak
7887 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7888 memory. This structure is used by the PKCS#7 and CMS routines so any
7889 application which reads PKCS#7 or CMS data from untrusted sources is
7890 affected. SSL/TLS is not affected.
7892 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7898 * Race condition handling PSK identify hint
7900 If PSK identity hints are received by a multi-threaded client then
7901 the values are wrongly updated in the parent SSL_CTX structure. This can
7902 result in a race condition potentially leading to a double free of the
7908 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7910 * Malformed ECParameters causes infinite loop
7912 When processing an ECParameters structure OpenSSL enters an infinite loop
7913 if the curve specified is over a specially malformed binary polynomial
7916 This can be used to perform denial of service against any
7917 system which processes public keys, certificate requests or
7918 certificates. This includes TLS clients and TLS servers with
7919 client authentication enabled.
7921 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7926 * Exploitable out-of-bounds read in X509_cmp_time
7928 X509_cmp_time does not properly check the length of the ASN1_TIME
7929 string and can read a few bytes out of bounds. In addition,
7930 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7933 An attacker can use this to craft malformed certificates and CRLs of
7934 various sizes and potentially cause a segmentation fault, resulting in
7935 a DoS on applications that verify certificates or CRLs. TLS clients
7936 that verify CRLs are affected. TLS clients and servers with client
7937 authentication enabled may be affected if they use custom verification
7940 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7941 independently by Hanno Böck.
7946 * PKCS7 crash with missing EnvelopedContent
7948 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7949 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7950 with missing content and trigger a NULL pointer dereference on parsing.
7952 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7953 structures from untrusted sources are affected. OpenSSL clients and
7954 servers are not affected.
7956 This issue was reported to OpenSSL by Michal Zalewski (Google).
7961 * CMS verify infinite loop with unknown hash function
7963 When verifying a signedData message the CMS code can enter an infinite loop
7964 if presented with an unknown hash function OID. This can be used to perform
7965 denial of service against any system which verifies signedData messages using
7967 This issue was reported to OpenSSL by Johannes Bauer.
7972 * Race condition handling NewSessionTicket
7974 If a NewSessionTicket is received by a multi-threaded client when attempting to
7975 reuse a previous ticket then a race condition can occur potentially leading to
7976 a double free of the ticket data.
7981 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7983 * Segmentation fault in ASN1_TYPE_cmp fix
7985 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7986 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7987 certificate signature algorithm consistency this can be used to crash any
7988 certificate verification operation and exploited in a DoS attack. Any
7989 application which performs certificate verification is vulnerable including
7990 OpenSSL clients and servers which enable client authentication.
7995 * ASN.1 structure reuse memory corruption fix
7997 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7998 memory corruption via an invalid write. Such reuse is and has been
7999 strongly discouraged and is believed to be rare.
8001 Applications that parse structures containing CHOICE or ANY DEFINED BY
8002 components may be affected. Certificate parsing (d2i_X509 and related
8003 functions) are however not affected. OpenSSL clients and servers are
8009 * PKCS7 NULL pointer dereferences fix
8011 The PKCS#7 parsing code does not handle missing outer ContentInfo
8012 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8013 missing content and trigger a NULL pointer dereference on parsing.
8015 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8016 otherwise parse PKCS#7 structures from untrusted sources are
8017 affected. OpenSSL clients and servers are not affected.
8019 This issue was reported to OpenSSL by Michal Zalewski (Google).
8024 * DoS via reachable assert in SSLv2 servers fix
8026 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8027 servers that both support SSLv2 and enable export cipher suites by sending
8028 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8030 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8031 (OpenSSL development team).
8036 * Use After Free following d2i_ECPrivatekey error fix
8038 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8039 could cause a use after free condition. This, in turn, could cause a double
8040 free in several private key parsing functions (such as d2i_PrivateKey
8041 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8042 for applications that receive EC private keys from untrusted
8043 sources. This scenario is considered rare.
8045 This issue was discovered by the BoringSSL project and fixed in their
8051 * X509_to_X509_REQ NULL pointer deref fix
8053 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8054 the certificate key is invalid. This function is rarely used in practice.
8056 This issue was discovered by Brian Carpenter.
8061 * Removed the export ciphers from the DEFAULT ciphers
8065 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8067 * Build fixes for the Windows and OpenVMS platforms
8069 *Matt Caswell and Richard Levitte*
8071 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8073 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8074 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8075 dereference. This could lead to a Denial Of Service attack. Thanks to
8076 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8081 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8082 dtls1_buffer_record function under certain conditions. In particular this
8083 could occur if an attacker sent repeated DTLS records with the same
8084 sequence number but for the next epoch. The memory leak could be exploited
8085 by an attacker in a Denial of Service attack through memory exhaustion.
8086 Thanks to Chris Mueller for reporting this issue.
8091 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8092 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8093 method would be set to NULL which could later result in a NULL pointer
8094 dereference. Thanks to Frank Schmirler for reporting this issue.
8099 * Abort handshake if server key exchange message is omitted for ephemeral
8102 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8103 reporting this issue.
8108 * Remove non-export ephemeral RSA code on client and server. This code
8109 violated the TLS standard by allowing the use of temporary RSA keys in
8110 non-export ciphersuites and could be used by a server to effectively
8111 downgrade the RSA key length used to a value smaller than the server
8112 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8113 INRIA or reporting this issue.
8118 * Fixed issue where DH client certificates are accepted without verification.
8119 An OpenSSL server will accept a DH certificate for client authentication
8120 without the certificate verify message. This effectively allows a client to
8121 authenticate without the use of a private key. This only affects servers
8122 which trust a client certificate authority which issues certificates
8123 containing DH keys: these are extremely rare and hardly ever encountered.
8124 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8130 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8131 results on some platforms, including x86_64. This bug occurs at random
8132 with a very low probability, and is not known to be exploitable in any
8133 way, though its exact impact is difficult to determine. Thanks to Pieter
8134 Wuille (Blockstream) who reported this issue and also suggested an initial
8135 fix. Further analysis was conducted by the OpenSSL development team and
8136 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8137 the OpenSSL core team.
8142 * Fix various certificate fingerprint issues.
8144 By using non-DER or invalid encodings outside the signed portion of a
8145 certificate the fingerprint can be changed without breaking the signature.
8146 Although no details of the signed portion of the certificate can be changed
8147 this can cause problems with some applications: e.g. those using the
8148 certificate fingerprint for blacklists.
8150 1. Reject signatures with non zero unused bits.
8152 If the BIT STRING containing the signature has non zero unused bits reject
8153 the signature. All current signature algorithms require zero unused bits.
8155 2. Check certificate algorithm consistency.
8157 Check the AlgorithmIdentifier inside TBS matches the one in the
8158 certificate signature. NB: this will result in signature failure
8159 errors for some broken certificates.
8161 Thanks to Konrad Kraszewski from Google for reporting this issue.
8163 3. Check DSA/ECDSA signatures use DER.
8165 Reencode DSA/ECDSA signatures and compare with the original received
8166 signature. Return an error if there is a mismatch.
8168 This will reject various cases including garbage after signature
8169 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8170 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8171 (negative or with leading zeroes).
8173 Further analysis was conducted and fixes were developed by Stephen Henson
8174 of the OpenSSL core team.
8180 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8182 * Session Ticket Memory Leak.
8184 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8185 integrity of that ticket is first verified. In the event of a session
8186 ticket integrity check failing, OpenSSL will fail to free memory
8187 causing a memory leak. By sending a large number of invalid session
8188 tickets an attacker could exploit this issue in a Denial Of Service
8194 * Build option no-ssl3 is incomplete.
8196 When OpenSSL is configured with "no-ssl3" as a build option, servers
8197 could accept and complete a SSL 3.0 handshake, and clients could be
8198 configured to send them.
8201 *Akamai and the OpenSSL team*
8203 * Add support for TLS_FALLBACK_SCSV.
8204 Client applications doing fallback retries should call
8205 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8208 *Adam Langley, Bodo Moeller*
8210 * Add additional DigestInfo checks.
8212 Reencode DigestInto in DER and check against the original when
8213 verifying RSA signature: this will reject any improperly encoded
8214 DigestInfo structures.
8216 Note: this is a precautionary measure and no attacks are currently known.
8220 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8222 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8223 to a denial of service attack. A malicious server can crash the client
8224 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8225 ciphersuite and sending carefully crafted handshake messages.
8227 Thanks to Felix Gröbert (Google) for discovering and researching this
8233 * By sending carefully crafted DTLS packets an attacker could cause openssl
8234 to leak memory. This can be exploited through a Denial of Service attack.
8235 Thanks to Adam Langley for discovering and researching this issue.
8240 * An attacker can force openssl to consume large amounts of memory whilst
8241 processing DTLS handshake messages. This can be exploited through a
8242 Denial of Service attack.
8243 Thanks to Adam Langley for discovering and researching this issue.
8248 * An attacker can force an error condition which causes openssl to crash
8249 whilst processing DTLS packets due to memory being freed twice. This
8250 can be exploited through a Denial of Service attack.
8251 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8257 * If a multithreaded client connects to a malicious server using a resumed
8258 session and the server sends an ec point format extension it could write
8259 up to 255 bytes to freed memory.
8261 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8267 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8268 X509_name_oneline, X509_name_print_ex et al. to leak some information
8269 from the stack. Applications may be affected if they echo pretty printing
8270 output to the attacker.
8272 Thanks to Ivan Fratric (Google) for discovering this issue.
8275 *Emilia Käsper, and Steve Henson*
8277 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8278 for corner cases. (Certain input points at infinity could lead to
8279 bogus results, with non-infinity inputs mapped to infinity too.)
8283 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8285 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8286 handshake can force the use of weak keying material in OpenSSL
8287 SSL/TLS clients and servers.
8289 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8290 researching this issue. ([CVE-2014-0224])
8292 *KIKUCHI Masashi, Steve Henson*
8294 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8295 OpenSSL DTLS client the code can be made to recurse eventually crashing
8298 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8301 *Imre Rad, Steve Henson*
8303 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8304 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8305 client or server. This is potentially exploitable to run arbitrary
8306 code on a vulnerable client or server.
8308 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8310 *Jüri Aedla, Steve Henson*
8312 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8313 are subject to a denial of service attack.
8315 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8316 this issue. ([CVE-2014-3470])
8318 *Felix Gröbert, Ivan Fratric, Steve Henson*
8320 * Harmonize version and its documentation. -f flag is used to display
8323 *mancha <mancha1@zoho.com>*
8325 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8326 in i2d_ECPrivateKey.
8328 *mancha <mancha1@zoho.com>*
8330 * Fix some double frees. These are not thought to be exploitable.
8332 *mancha <mancha1@zoho.com>*
8334 * Fix for the attack described in the paper "Recovering OpenSSL
8335 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8336 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8337 <http://eprint.iacr.org/2014/140>
8339 Thanks to Yuval Yarom and Naomi Benger for discovering this
8340 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8342 *Yuval Yarom and Naomi Benger*
8344 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8346 * Keep original DTLS digest and encryption contexts in retransmission
8347 structures so we can use the previous session parameters if they need
8348 to be resent. ([CVE-2013-6450])
8352 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8353 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8354 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8355 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8356 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8357 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8359 *Rob Stradling, Adam Langley*
8361 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8363 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8365 This addresses the flaw in CBC record processing discovered by
8366 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8367 at: <http://www.isg.rhul.ac.uk/tls/>
8369 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8370 Security Group at Royal Holloway, University of London
8371 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8372 Emilia Käsper for the initial patch.
8375 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8377 * Return an error when checking OCSP signatures when key is NULL.
8378 This fixes a DoS attack. ([CVE-2013-0166])
8382 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8383 the right response is stapled. Also change SSL_get_certificate()
8384 so it returns the certificate actually sent.
8385 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8386 (This is a backport)
8388 *Rob Stradling <rob.stradling@comodo.com>*
8390 * Fix possible deadlock when decoding public keys.
8394 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8396 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8399 * Sanity check record length before skipping explicit IV in DTLS
8402 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8403 fuzzing as a service testing platform.
8408 * Initialise tkeylen properly when encrypting CMS messages.
8409 Thanks to Solar Designer of Openwall for reporting this issue.
8413 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8415 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8416 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8417 in CRYPTO_realloc_clean.
8419 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8420 issue and to Adam Langley <agl@chromium.org> for fixing it.
8423 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8425 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8427 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8428 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8429 content decryption and always return the same error. Note: this attack
8430 needs on average 2^20 messages so it only affects automated senders. The
8431 old behaviour can be re-enabled in the CMS code by setting the
8432 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8433 an MMA defence is not necessary.
8434 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8435 this issue. ([CVE-2012-0884])
8439 * Fix CVE-2011-4619: make sure we really are receiving a
8440 client hello before rejecting multiple SGC restarts. Thanks to
8441 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8445 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8447 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8448 Thanks to Antonio Martin, Enterprise Secure Access Research and
8449 Development, Cisco Systems, Inc. for discovering this bug and
8450 preparing a fix. ([CVE-2012-0050])
8454 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8456 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8457 of the Vaudenay padding oracle attack on CBC mode encryption
8458 which enables an efficient plaintext recovery attack against
8459 the OpenSSL implementation of DTLS. Their attack exploits timing
8460 differences arising during decryption processing. A research
8461 paper describing this attack can be found at:
8462 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8463 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8464 Security Group at Royal Holloway, University of London
8465 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8466 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8467 for preparing the fix. ([CVE-2011-4108])
8469 *Robin Seggelmann, Michael Tuexen*
8471 * Clear bytes used for block padding of SSL 3.0 records.
8474 *Adam Langley (Google)*
8476 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8477 Kadianakis <desnacked@gmail.com> for discovering this issue and
8478 Adam Langley for preparing the fix. ([CVE-2011-4619])
8480 *Adam Langley (Google)*
8482 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8484 *Andrey Kulikov <amdeich@gmail.com>*
8486 * Prevent malformed RFC3779 data triggering an assertion failure.
8487 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8488 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8490 *Rob Austein <sra@hactrn.net>*
8492 * Improved PRNG seeding for VOS.
8494 *Paul Green <Paul.Green@stratus.com>*
8496 * Fix ssl_ciph.c set-up race.
8498 *Adam Langley (Google)*
8500 * Fix spurious failures in ecdsatest.c.
8502 *Emilia Käsper (Google)*
8504 * Fix the BIO_f_buffer() implementation (which was mixing different
8505 interpretations of the `..._len` fields).
8507 *Adam Langley (Google)*
8509 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8510 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8511 threads won't reuse the same blinding coefficients.
8513 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8514 lock to call BN_BLINDING_invert_ex, and avoids one use of
8515 BN_BLINDING_update for each BN_BLINDING structure (previously,
8516 the last update always remained unused).
8518 *Emilia Käsper (Google)*
8520 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8522 *Bob Buckholz (Google)*
8524 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8526 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8527 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8529 *Kaspar Brand <ossl@velox.ch>*
8531 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8532 for multi-threaded use of ECDH. ([CVE-2011-3210])
8534 *Adam Langley (Google)*
8536 * Fix x509_name_ex_d2i memory leak on bad inputs.
8540 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8541 signature public key algorithm by using OID xref utilities instead.
8542 Before this you could only use some ECC ciphersuites with SHA1 only.
8546 * Add protection against ECDSA timing attacks as mentioned in the paper
8547 by Billy Bob Brumley and Nicola Tuveri, see:
8548 <http://eprint.iacr.org/2011/232.pdf>
8550 *Billy Bob Brumley and Nicola Tuveri*
8552 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8554 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8556 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8558 * Fix bug in string printing code: if *any* escaping is enabled we must
8559 escape the escape character (backslash) or the resulting string is
8564 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8566 * Disable code workaround for ancient and obsolete Netscape browsers
8567 and servers: an attacker can use it in a ciphersuite downgrade attack.
8568 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8572 * Fixed J-PAKE implementation error, originally discovered by
8573 Sebastien Martini, further info and confirmation from Stefan
8574 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8578 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8580 * Fix extension code to avoid race conditions which can result in a buffer
8581 overrun vulnerability: resumed sessions must not be modified as they can
8582 be shared by multiple threads. CVE-2010-3864
8586 * Fix WIN32 build system to correctly link an ENGINE directory into
8591 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8593 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8596 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8598 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8600 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8601 context. The operation can be customised via the ctrl mechanism in
8602 case ENGINEs want to include additional functionality.
8606 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8610 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8611 output hashes compatible with older versions of OpenSSL.
8613 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8615 * Fix compression algorithm handling: if resuming a session use the
8616 compression algorithm of the resumed session instead of determining
8617 it from client hello again. Don't allow server to change algorithm.
8621 * Add load_crls() function to commands tidying load_certs() too. Add option
8622 to verify utility to allow additional CRLs to be included.
8626 * Update OCSP request code to permit adding custom headers to the request:
8627 some responders need this.
8631 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8634 *Julia Lawall <julia@diku.dk>*
8636 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8637 needlessly dereferenced structures, used obsolete functions and
8638 didn't handle all updated verify codes correctly.
8642 * Disable MD2 in the default configuration.
8646 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8647 indicate the initial BIO being pushed or popped. This makes it possible
8648 to determine whether the BIO is the one explicitly called or as a result
8649 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8650 it handles reference counts correctly and doesn't zero out the I/O bio
8651 when it is not being explicitly popped. WARNING: applications which
8652 included workarounds for the old buggy behaviour will need to be modified
8653 or they could free up already freed BIOs.
8657 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8658 renaming to all platforms (within the 0.9.8 branch, this was
8659 done conditionally on Netware platforms to avoid a name clash).
8661 *Guenter <lists@gknw.net>*
8663 * Add ECDHE and PSK support to DTLS.
8665 *Michael Tuexen <tuexen@fh-muenster.de>*
8667 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8672 * Add "missing" function EVP_MD_flags() (without this the only way to
8673 retrieve a digest flags is by accessing the structure directly. Update
8674 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8675 or cipher is registered as in the "from" argument. Print out all
8676 registered digests in the dgst usage message instead of manually
8677 attempting to work them out.
8681 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8682 this allows the use of compression and extensions. Change default cipher
8683 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8684 by default unless an application cipher string requests it.
8688 * Alter match criteria in PKCS12_parse(). It used to try to use local
8689 key ids to find matching certificates and keys but some PKCS#12 files
8690 don't follow the (somewhat unwritten) rules and this strategy fails.
8691 Now just gather all certificates together and the first private key
8692 then look for the first certificate that matches the key.
8696 * Support use of registered digest and cipher names for dgst and cipher
8697 commands instead of having to add each one as a special case. So now
8704 openssl dgst -sha256 foo
8706 and this works for ENGINE based algorithms too.
8710 * Update Gost ENGINE to support parameter files.
8712 *Victor B. Wagner <vitus@cryptocom.ru>*
8714 * Support GeneralizedTime in ca utility.
8716 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8718 * Enhance the hash format used for certificate directory links. The new
8719 form uses the canonical encoding (meaning equivalent names will work
8720 even if they aren't identical) and uses SHA1 instead of MD5. This form
8721 is incompatible with the older format and as a result c_rehash should
8722 be used to rebuild symbolic links.
8726 * Make PKCS#8 the default write format for private keys, replacing the
8727 traditional format. This form is standardised, more secure and doesn't
8728 include an implicit MD5 dependency.
8732 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8733 committed to OpenSSL should pass this lot as a minimum.
8737 * Add session ticket override functionality for use by EAP-FAST.
8739 *Jouni Malinen <j@w1.fi>*
8741 * Modify HMAC functions to return a value. Since these can be implemented
8742 in an ENGINE errors can occur.
8746 * Type-checked OBJ_bsearch_ex.
8750 * Type-checked OBJ_bsearch. Also some constification necessitated
8751 by type-checking. Still to come: TXT_DB, bsearch(?),
8752 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8757 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8758 seconds to a tm structure directly, instead of going through OS
8759 specific date routines. This avoids any issues with OS routines such
8760 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8761 and X509_time_adj_ex() to cover the extended range. The existing
8762 X509_time_adj() is still usable and will no longer have any date issues.
8766 * Delta CRL support. New use deltas option which will attempt to locate
8767 and search any appropriate delta CRLs available.
8769 This work was sponsored by Google.
8773 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8774 code and add additional score elements. Validate alternate CRL paths
8775 as part of the CRL checking and indicate a new error "CRL path validation
8776 error" in this case. Applications wanting additional details can use
8777 the verify callback and check the new "parent" field. If this is not
8778 NULL CRL path validation is taking place. Existing applications won't
8779 see this because it requires extended CRL support which is off by
8782 This work was sponsored by Google.
8786 * Support for freshest CRL extension.
8788 This work was sponsored by Google.
8792 * Initial indirect CRL support. Currently only supported in the CRLs
8793 passed directly and not via lookup. Process certificate issuer
8794 CRL entry extension and lookup CRL entries by bother issuer name
8795 and serial number. Check and process CRL issuer entry in IDP extension.
8797 This work was sponsored by Google.
8801 * Add support for distinct certificate and CRL paths. The CRL issuer
8802 certificate is validated separately in this case. Only enabled if
8803 an extended CRL support flag is set: this flag will enable additional
8804 CRL functionality in future.
8806 This work was sponsored by Google.
8810 * Add support for policy mappings extension.
8812 This work was sponsored by Google.
8816 * Fixes to pathlength constraint, self issued certificate handling,
8817 policy processing to align with RFC3280 and PKITS tests.
8819 This work was sponsored by Google.
8823 * Support for name constraints certificate extension. DN, email, DNS
8824 and URI types are currently supported.
8826 This work was sponsored by Google.
8830 * To cater for systems that provide a pointer-based thread ID rather
8831 than numeric, deprecate the current numeric thread ID mechanism and
8832 replace it with a structure and associated callback type. This
8833 mechanism allows a numeric "hash" to be extracted from a thread ID in
8834 either case, and on platforms where pointers are larger than 'long',
8835 mixing is done to help ensure the numeric 'hash' is usable even if it
8836 can't be guaranteed unique. The default mechanism is to use "&errno"
8837 as a pointer-based thread ID to distinguish between threads.
8839 Applications that want to provide their own thread IDs should now use
8840 CRYPTO_THREADID_set_callback() to register a callback that will call
8841 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8843 Note that ERR_remove_state() is now deprecated, because it is tied
8844 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8845 to free the current thread's error state should be replaced by
8846 ERR_remove_thread_state(NULL).
8848 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8849 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8850 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8851 application was previously providing a numeric thread callback that
8852 was inappropriate for distinguishing threads, then uniqueness might
8853 have been obtained with &errno that happened immediately in the
8854 intermediate development versions of OpenSSL; this is no longer the
8855 case, the numeric thread callback will now override the automatic use
8858 *Geoff Thorpe, with help from Bodo Moeller*
8860 * Initial support for different CRL issuing certificates. This covers a
8861 simple case where the self issued certificates in the chain exist and
8862 the real CRL issuer is higher in the existing chain.
8864 This work was sponsored by Google.
8868 * Removed effectively defunct crypto/store from the build.
8872 * Revamp of STACK to provide stronger type-checking. Still to come:
8873 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8874 ASN1_STRING, CONF_VALUE.
8878 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8879 RAM on SSL connections. This option can save about 34k per idle SSL.
8883 * Revamp of LHASH to provide stronger type-checking. Still to come:
8884 STACK, TXT_DB, bsearch, qsort.
8888 * Initial support for Cryptographic Message Syntax (aka CMS) based
8889 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8890 support for data, signedData, compressedData, digestedData and
8891 encryptedData, envelopedData types included. Scripts to check against
8892 RFC4134 examples draft and interop and consistency checks of many
8893 content types and variants.
8897 * Add options to enc utility to support use of zlib compression BIO.
8901 * Extend mk1mf to support importing of options and assembly language
8902 files from Configure script, currently only included in VC-WIN32.
8903 The assembly language rules can now optionally generate the source
8904 files from the associated perl scripts.
8908 * Implement remaining functionality needed to support GOST ciphersuites.
8909 Interop testing has been performed using CryptoPro implementations.
8911 *Victor B. Wagner <vitus@cryptocom.ru>*
8913 * s390x assembler pack.
8917 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8922 * Implement Opaque PRF Input TLS extension as specified in
8923 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8924 official specification yet and no extension type assignment by
8925 IANA exists, this extension (for now) will have to be explicitly
8926 enabled when building OpenSSL by providing the extension number
8927 to use. For example, specify an option
8929 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8931 to the "config" or "Configure" script to enable the extension,
8932 assuming extension number 0x9527 (which is a completely arbitrary
8933 and unofficial assignment based on the MD5 hash of the Internet
8934 Draft). Note that by doing so, you potentially lose
8935 interoperability with other TLS implementations since these might
8936 be using the same extension number for other purposes.
8938 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8939 opaque PRF input value to use in the handshake. This will create
8940 an internal copy of the length-'len' string at 'src', and will
8941 return non-zero for success.
8943 To get more control and flexibility, provide a callback function
8946 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8947 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8951 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8954 Callback function 'cb' will be called in handshakes, and is
8955 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8956 Argument 'arg' is for application purposes (the value as given to
8957 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8958 be provided to the callback function). The callback function
8959 has to return non-zero to report success: usually 1 to use opaque
8960 PRF input just if possible, or 2 to enforce use of the opaque PRF
8961 input. In the latter case, the library will abort the handshake
8962 if opaque PRF input is not successfully negotiated.
8964 Arguments 'peerinput' and 'len' given to the callback function
8965 will always be NULL and 0 in the case of a client. A server will
8966 see the client's opaque PRF input through these variables if
8967 available (NULL and 0 otherwise). Note that if the server
8968 provides an opaque PRF input, the length must be the same as the
8969 length of the client's opaque PRF input.
8971 Note that the callback function will only be called when creating
8972 a new session (session resumption can resume whatever was
8973 previously negotiated), and will not be called in SSL 2.0
8974 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8975 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8976 for applications that need to enforce opaque PRF input.
8980 * Update ssl code to support digests other than SHA1+MD5 for handshake
8983 *Victor B. Wagner <vitus@cryptocom.ru>*
8985 * Add RFC4507 support to OpenSSL. This includes the corrections in
8986 RFC4507bis. The encrypted ticket format is an encrypted encoded
8987 SSL_SESSION structure, that way new session features are automatically
8990 If a client application caches session in an SSL_SESSION structure
8991 support is transparent because tickets are now stored in the encoded
8994 The SSL_CTX structure automatically generates keys for ticket
8995 protection in servers so again support should be possible
8996 with no application modification.
8998 If a client or server wishes to disable RFC4507 support then the option
8999 SSL_OP_NO_TICKET can be set.
9001 Add a TLS extension debugging callback to allow the contents of any client
9002 or server extensions to be examined.
9004 This work was sponsored by Google.
9008 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9009 OpenSSL should now compile cleanly on gcc 4.2
9011 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9013 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9014 support including streaming MAC support: this is required for GOST
9015 ciphersuite support.
9017 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9019 * Add option -stream to use PKCS#7 streaming in smime utility. New
9020 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9021 to output in BER and PEM format.
9025 * Experimental support for use of HMAC via EVP_PKEY interface. This
9026 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9027 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9028 ENGINE support for HMAC keys which are unextractable. New -mac and
9029 -macopt options to dgst utility.
9033 * New option -sigopt to dgst utility. Update dgst to use
9034 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9035 alternative signing parameters such as X9.31 or PSS in the dgst
9040 * Change ssl_cipher_apply_rule(), the internal function that does
9041 the work each time a ciphersuite string requests enabling
9042 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9043 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9044 the order of disabled ciphersuites such that those ciphersuites
9045 that most recently went from enabled to disabled not only stay
9046 in order with respect to each other, but also have higher priority
9047 than other disabled ciphersuites the next time ciphersuites are
9050 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9051 the same ciphersuites as with "HIGH" alone, but in a specific
9052 order where the PSK ciphersuites come first (since they are the
9053 most recently disabled ciphersuites when "HIGH" is parsed).
9055 Also, change ssl_create_cipher_list() (using this new
9056 functionality) such that between otherwise identical
9057 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9062 * Change ssl_create_cipher_list() so that it automatically
9063 arranges the ciphersuites in reasonable order before starting
9064 to process the rule string. Thus, the definition for "DEFAULT"
9065 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9066 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9067 This makes it much easier to arrive at a reasonable default order
9068 in applications for which anonymous ciphers are OK (meaning
9069 that you can't actually use DEFAULT).
9071 *Bodo Moeller; suggested by Victor Duchovni*
9073 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9074 processing) into multiple integers instead of setting
9075 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9076 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9077 (These masks as well as the individual bit definitions are hidden
9078 away into the non-exported interface ssl/ssl_locl.h, so this
9079 change to the definition of the SSL_CIPHER structure shouldn't
9080 affect applications.) This give us more bits for each of these
9081 categories, so there is no longer a need to coagulate AES128 and
9082 AES256 into a single algorithm bit, and to coagulate Camellia128
9083 and Camellia256 into a single algorithm bit, which has led to all
9086 Thus, among other things, the kludge introduced in 0.9.7m and
9087 0.9.8e for masking out AES256 independently of AES128 or masking
9088 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9090 With the change, we also introduce new ciphersuite aliases that
9091 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9096 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9097 Use the leftmost N bytes of the signature input if the input is
9098 larger than the prime q (with N being the size in bytes of q).
9102 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9103 it yet and it is largely untested.
9107 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9111 * Initial incomplete changes to avoid need for function casts in OpenSSL
9112 some compilers (gcc 4.2 and later) reject their use. Safestack is
9113 reimplemented. Update ASN1 to avoid use of legacy functions.
9117 * Win32/64 targets are linked with Winsock2.
9121 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9122 to external functions. This can be used to increase CRL handling
9123 efficiency especially when CRLs are very large by (for example) storing
9124 the CRL revoked certificates in a database.
9128 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9129 new CRLs added to a directory can be used. New command line option
9130 -verify_return_error to s_client and s_server. This causes real errors
9131 to be returned by the verify callback instead of carrying on no matter
9132 what. This reflects the way a "real world" verify callback would behave.
9136 * GOST engine, supporting several GOST algorithms and public key formats.
9137 Kindly donated by Cryptocom.
9141 * Partial support for Issuing Distribution Point CRL extension. CRLs
9142 partitioned by DP are handled but no indirect CRL or reason partitioning
9143 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9144 selected via a scoring technique which handles IDP and AKID in CRLs.
9148 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9149 will ultimately be used for all verify operations: this will remove the
9150 X509_STORE dependency on certificate verification and allow alternative
9151 lookup methods. X509_STORE based implementations of these two callbacks.
9155 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9156 Modify get_crl() to find a valid (unexpired) CRL if possible.
9160 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9161 this would be called X509_CRL_cmp() but that name is already used by
9162 a function that just compares CRL issuer names. Cache several CRL
9163 extensions in X509_CRL structure and cache CRLDP in X509.
9167 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9168 this maps equivalent X509_NAME structures into a consistent structure.
9169 Name comparison can then be performed rapidly using memcmp().
9173 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9178 * Allow digests to supply their own micalg string for S/MIME type using
9179 the ctrl EVP_MD_CTRL_MICALG.
9183 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9184 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9185 ctrl. It can then customise the structure before and/or after signing
9190 * New function OBJ_add_sigid() to allow application defined signature OIDs
9191 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9192 to free up any added signature OIDs.
9196 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9197 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9198 digest and cipher tables. New options added to openssl utility:
9199 list-message-digest-algorithms and list-cipher-algorithms.
9203 * Change the array representation of binary polynomials: the list
9204 of degrees of non-zero coefficients is now terminated with -1.
9205 Previously it was terminated with 0, which was also part of the
9206 value; thus, the array representation was not applicable to
9207 polynomials where t^0 has coefficient zero. This change makes
9208 the array representation useful in a more general context.
9212 * Various modifications and fixes to SSL/TLS cipher string
9213 handling. For ECC, the code now distinguishes between fixed ECDH
9214 with RSA certificates on the one hand and with ECDSA certificates
9215 on the other hand, since these are separate ciphersuites. The
9216 unused code for Fortezza ciphersuites has been removed.
9218 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9219 (not "ECDHE"). For consistency with the code for DH
9220 certificates, use of ECDH certificates is now considered ECDH
9221 authentication, not RSA or ECDSA authentication (the latter is
9222 merely the CA's signing algorithm and not actively used in the
9225 The temporary ciphersuite alias "ECCdraft" is no longer
9226 available, and ECC ciphersuites are no longer excluded from "ALL"
9227 and "DEFAULT". The following aliases now exist for RFC 4492
9228 ciphersuites, most of these by analogy with the DH case:
9230 kECDHr - ECDH cert, signed with RSA
9231 kECDHe - ECDH cert, signed with ECDSA
9232 kECDH - ECDH cert (signed with either RSA or ECDSA)
9233 kEECDH - ephemeral ECDH
9234 ECDH - ECDH cert or ephemeral ECDH
9240 AECDH - anonymous ECDH
9241 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9245 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9246 Use correct micalg parameters depending on digest(s) in signed message.
9250 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9251 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9255 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9256 an engine to register a method. Add ENGINE lookups for methods and
9257 functional reference processing.
9261 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9262 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9267 * New -resign option to smime utility. This adds one or more signers
9268 to an existing PKCS#7 signedData structure. Also -md option to use an
9269 alternative message digest algorithm for signing.
9273 * Tidy up PKCS#7 routines and add new functions to make it easier to
9274 create PKCS7 structures containing multiple signers. Update smime
9275 application to support multiple signers.
9279 * New -macalg option to pkcs12 utility to allow setting of an alternative
9284 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9285 Reorganize PBE internals to lookup from a static table using NIDs,
9286 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9287 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9288 PRF which will be automatically used with PBES2.
9292 * Replace the algorithm specific calls to generate keys in "req" with the
9297 * Update PKCS#7 enveloped data routines to use new API. This is now
9298 supported by any public key method supporting the encrypt operation. A
9299 ctrl is added to allow the public key algorithm to examine or modify
9300 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9305 * Add a ctrl to asn1 method to allow a public key algorithm to express
9306 a default digest type to use. In most cases this will be SHA1 but some
9307 algorithms (such as GOST) need to specify an alternative digest. The
9308 return value indicates how strong the preference is 1 means optional and
9309 2 is mandatory (that is it is the only supported type). Modify
9310 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9311 use the default md. Update openssl utilities to use the default digest
9312 type for signing if it is not explicitly indicated.
9316 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9317 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9318 signing method from the key type. This effectively removes the link
9319 between digests and public key types.
9323 * Add an OID cross reference table and utility functions. Its purpose is to
9324 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9325 rsaEncryption. This will allow some of the algorithm specific hackery
9326 needed to use the correct OID to be removed.
9330 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9331 structures for PKCS7_sign(). They are now set up by the relevant public
9336 * Add provisional EC pkey method with support for ECDSA and ECDH.
9340 * Add support for key derivation (agreement) in the API, DH method and
9345 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9346 public and private key formats. As a side effect these add additional
9347 command line functionality not previously available: DSA signatures can be
9348 generated and verified using pkeyutl and DH key support and generation in
9355 *Oliver Tappe <zooey@hirschkaefer.de>*
9357 * New make target "install_html_docs" installs HTML renditions of the
9360 *Oliver Tappe <zooey@hirschkaefer.de>*
9362 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9363 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9364 support key and parameter generation and add initial key generation
9365 functionality for RSA.
9369 * Add functions for main EVP_PKEY_method operations. The undocumented
9370 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9371 `EVP_PKEY_{encrypt,decrypt}_old`.
9375 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9376 key API, doesn't do much yet.
9380 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9381 public key algorithms. New option to openssl utility:
9382 "list-public-key-algorithms" to print out info.
9386 * Implement the Supported Elliptic Curves Extension for
9387 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9391 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9392 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9396 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9397 utilities such as rsa, dsa, dsaparam etc except they process any key
9402 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9403 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9404 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9409 * Initial support for pluggable public key ASN1.
9410 De-spaghettify the public key ASN1 handling. Move public and private
9411 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9412 algorithm specific handling to a single module within the relevant
9413 algorithm directory. Add functions to allow (near) opaque processing
9414 of public and private key structures.
9418 * Implement the Supported Point Formats Extension for
9419 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9423 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9424 for the psk identity [hint] and the psk callback functions to the
9425 SSL_SESSION, SSL and SSL_CTX structure.
9428 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9432 SSL_CTX_use_psk_identity_hint
9433 SSL_get_psk_identity_hint
9434 SSL_get_psk_identity
9435 SSL_use_psk_identity_hint
9437 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9439 * Add RFC 3161 compliant time stamp request creation, response generation
9440 and response verification functionality.
9442 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9444 * Add initial support for TLS extensions, specifically for the server_name
9445 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9446 have new members for a host name. The SSL data structure has an
9447 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9448 stored in that context to allow for session resumption, even after the
9449 SSL has been switched to a new SSL_CTX in reaction to a client's
9450 server_name extension.
9452 New functions (subject to change):
9454 SSL_get_servername()
9455 SSL_get_servername_type()
9458 New CTRL codes and macros (subject to change):
9460 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9461 - SSL_CTX_set_tlsext_servername_callback()
9462 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9463 - SSL_CTX_set_tlsext_servername_arg()
9464 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9466 openssl s_client has a new '-servername ...' option.
9468 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9469 '-key2 ...', '-servername_fatal' (subject to change). This allows
9470 testing the HostName extension for a specific single host name ('-cert'
9471 and '-key' remain fallbacks for handshakes without HostName
9472 negotiation). If the unrecognized_name alert has to be sent, this by
9473 default is a warning; it becomes fatal with the '-servername_fatal'
9476 *Peter Sylvester, Remy Allais, Christophe Renou*
9478 * Whirlpool hash implementation is added.
9482 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9483 bn(64,32). Because of instruction set limitations it doesn't have
9484 any negative impact on performance. This was done mostly in order
9485 to make it possible to share assembler modules, such as bn_mul_mont
9486 implementations, between 32- and 64-bit builds without hassle.
9490 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9491 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9496 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9497 dedicated Montgomery multiplication procedure, is introduced.
9498 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9499 "64-bit" performance on certain 32-bit targets.
9503 * New option SSL_OP_NO_COMP to disable use of compression selectively
9504 in SSL structures. New SSL ctrl to set maximum send fragment size.
9505 Save memory by setting the I/O buffer sizes dynamically instead of
9506 using the maximum available value.
9510 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9511 in addition to the text details.
9515 * Very, very preliminary EXPERIMENTAL support for printing of general
9516 ASN1 structures. This currently produces rather ugly output and doesn't
9517 handle several customised structures at all.
9521 * Integrated support for PVK file format and some related formats such
9522 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9523 these in the 'rsa' and 'dsa' utilities.
9527 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9531 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9532 place for the (very old) "NETSCAPE" format certificates which are now
9533 handled using new ASN1 code equivalents.
9537 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9538 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9539 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9543 * Modify CRL distribution points extension code to print out previously
9544 unsupported fields. Enhance extension setting code to allow setting of
9549 * Add print and set support for Issuing Distribution Point CRL extension.
9553 * Change 'Configure' script to enable Camellia by default.
9560 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9562 * When rejecting SSL/TLS records due to an incorrect version number, never
9563 update s->server with a new major version number. As of
9564 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9565 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9566 the previous behavior could result in a read attempt at NULL when
9567 receiving specific incorrect SSL/TLS records once record payload
9568 protection is active. ([CVE-2010-0740])
9570 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9572 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9573 could be crashed if the relevant tables were not present (e.g. chrooted).
9575 *Tomas Hoger <thoger@redhat.com>*
9577 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9579 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9581 *Martin Olsson, Neel Mehta*
9583 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9584 accommodate for stack sorting, always a write lock!).
9588 * On some versions of WIN32 Heap32Next is very slow. This can cause
9589 excessive delays in the RAND_poll(): over a minute. As a workaround
9590 include a time check in the inner Heap32Next loop too.
9594 * The code that handled flushing of data in SSL/TLS originally used the
9595 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9596 the problem outlined in PR#1949. The fix suggested there however can
9597 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9598 of Apache). So instead simplify the code to flush unconditionally.
9599 This should be fine since flushing with no data to flush is a no op.
9603 * Handle TLS versions 2.0 and later properly and correctly use the
9604 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9605 off ancient servers have a habit of sticking around for a while...
9609 * Modify compression code so it frees up structures without using the
9610 ex_data callbacks. This works around a problem where some applications
9611 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9612 restarting) then use compression (e.g. SSL with compression) later.
9613 This results in significant per-connection memory leaks and
9614 has caused some security issues including CVE-2008-1678 and
9619 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9620 change when encrypting or decrypting.
9624 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9625 connect and renegotiate with servers which do not support RI.
9626 Until RI is more widely deployed this option is enabled by default.
9630 * Add "missing" ssl ctrls to clear options and mode.
9634 * If client attempts to renegotiate and doesn't support RI respond with
9635 a no_renegotiation alert as required by RFC5746. Some renegotiating
9636 TLS clients will continue a connection gracefully when they receive
9637 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9638 waiting for a server hello which it will never receive. Now we treat a
9639 received no_renegotiation alert as a fatal error. This is because
9640 applications requesting a renegotiation might well expect it to succeed
9641 and would have no code in place to handle the server denying it so the
9642 only safe thing to do is to terminate the connection.
9646 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9647 peer supports secure renegotiation and 0 otherwise. Print out peer
9648 renegotiation support in s_client/s_server.
9652 * Replace the highly broken and deprecated SPKAC certification method with
9653 the updated NID creation version. This should correctly handle UTF8.
9657 * Implement RFC5746. Re-enable renegotiation but require the extension
9658 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9659 turns out to be a bad idea. It has been replaced by
9660 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9661 SSL_CTX_set_options(). This is really not recommended unless you
9662 know what you are doing.
9664 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9666 * Fixes to stateless session resumption handling. Use initial_ctx when
9667 issuing and attempting to decrypt tickets in case it has changed during
9668 servername handling. Use a non-zero length session ID when attempting
9669 stateless session resumption: this makes it possible to determine if
9670 a resumption has occurred immediately after receiving server hello
9671 (several places in OpenSSL subtly assume this) instead of later in
9676 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9677 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9678 fixes for a few places where the return code is not checked
9681 *Julia Lawall <julia@diku.dk>*
9683 * Add --strict-warnings option to Configure script to include devteam
9684 warnings in other configurations.
9688 * Add support for --libdir option and LIBDIR variable in makefiles. This
9689 makes it possible to install openssl libraries in locations which
9690 have names other than "lib", for example "/usr/lib64" which some
9693 *Steve Henson, based on patch from Jeremy Utley*
9695 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9696 X690 8.9.12 and can produce some misleading textual output of OIDs.
9698 *Steve Henson, reported by Dan Kaminsky*
9700 * Delete MD2 from algorithm tables. This follows the recommendation in
9701 several standards that it is not used in new applications due to
9702 several cryptographic weaknesses. For binary compatibility reasons
9703 the MD2 API is still compiled in by default.
9707 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9712 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9713 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9716 *Guenter <lists@gknw.net>*
9718 * Fix the server certificate chain building code to use X509_verify_cert(),
9719 it used to have an ad-hoc builder which was unable to cope with anything
9720 other than a simple chain.
9722 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9724 * Don't check self signed certificate signatures in X509_verify_cert()
9725 by default (a flag can override this): it just wastes time without
9726 adding any security. As a useful side effect self signed root CAs
9727 with non-FIPS digests are now usable in FIPS mode.
9731 * In dtls1_process_out_of_seq_message() the check if the current message
9732 is already buffered was missing. For every new message was memory
9733 allocated, allowing an attacker to perform an denial of service attack
9734 with sending out of seq handshake messages until there is no memory
9735 left. Additionally every future message was buffered, even if the
9736 sequence number made no sense and would be part of another handshake.
9737 So only messages with sequence numbers less than 10 in advance will be
9738 buffered. ([CVE-2009-1378])
9740 *Robin Seggelmann, discovered by Daniel Mentz*
9742 * Records are buffered if they arrive with a future epoch to be
9743 processed after finishing the corresponding handshake. There is
9744 currently no limitation to this buffer allowing an attacker to perform
9745 a DOS attack with sending records with future epochs until there is no
9746 memory left. This patch adds the pqueue_size() function to determine
9747 the size of a buffer and limits the record buffer to 100 entries.
9750 *Robin Seggelmann, discovered by Daniel Mentz*
9752 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9753 parent structure is freed. ([CVE-2009-1379])
9757 * Handle non-blocking I/O properly in SSL_shutdown() call.
9759 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9761 * Add `2.5.4.*` OIDs
9763 *Ilya O. <vrghost@gmail.com>*
9765 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9767 * Disable renegotiation completely - this fixes a severe security
9768 problem ([CVE-2009-3555]) at the cost of breaking all
9769 renegotiation. Renegotiation can be re-enabled by setting
9770 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9771 run-time. This is really not recommended unless you know what
9776 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9778 * Don't set val to NULL when freeing up structures, it is freed up by
9779 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9780 zeroing past the valid field. ([CVE-2009-0789])
9782 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9784 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9785 checked correctly. This would allow some invalid signed attributes to
9786 appear to verify correctly. ([CVE-2009-0591])
9788 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9790 * Reject UniversalString and BMPString types with invalid lengths. This
9791 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9792 a legal length. ([CVE-2009-0590])
9796 * Set S/MIME signing as the default purpose rather than setting it
9797 unconditionally. This allows applications to override it at the store
9802 * Permit restricted recursion of ASN1 strings. This is needed in practice
9803 to handle some structures.
9807 * Improve efficiency of mem_gets: don't search whole buffer each time
9810 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9812 * New -hex option for openssl rand.
9816 * Print out UTF8String and NumericString when parsing ASN1.
9820 * Support NumericString type for name components.
9824 * Allow CC in the environment to override the automatically chosen
9825 compiler. Note that nothing is done to ensure flags work with the
9830 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9832 * Properly check EVP_VerifyFinal() and similar return values
9835 *Ben Laurie, Bodo Moeller, Google Security Team*
9837 * Enable TLS extensions by default.
9841 * Allow the CHIL engine to be loaded, whether the application is
9842 multithreaded or not. (This does not release the developer from the
9843 obligation to set up the dynamic locking callbacks.)
9845 *Sander Temme <sander@temme.net>*
9847 * Use correct exit code if there is an error in dgst command.
9849 *Steve Henson; problem pointed out by Roland Dirlewanger*
9851 * Tweak Configure so that you need to say "experimental-jpake" to enable
9852 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9856 * Add experimental JPAKE support, including demo authentication in
9857 s_client and s_server.
9861 * Set the comparison function in v3_addr_canonize().
9863 *Rob Austein <sra@hactrn.net>*
9865 * Add support for XMPP STARTTLS in s_client.
9867 *Philip Paeps <philip@freebsd.org>*
9869 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9870 to ensure that even with this option, only ciphersuites in the
9871 server's preference list will be accepted. (Note that the option
9872 applies only when resuming a session, so the earlier behavior was
9873 just about the algorithm choice for symmetric cryptography.)
9877 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9879 * Fix NULL pointer dereference if a DTLS server received
9880 ChangeCipherSpec as first record ([CVE-2009-1386]).
9884 * Fix a state transition in s3_srvr.c and d1_srvr.c
9885 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9889 * The fix in 0.9.8c that supposedly got rid of unsafe
9890 double-checked locking was incomplete for RSA blinding,
9891 addressing just one layer of what turns out to have been
9892 doubly unsafe triple-checked locking.
9894 So now fix this for real by retiring the MONT_HELPER macro
9895 in crypto/rsa/rsa_eay.c.
9897 *Bodo Moeller; problem pointed out by Marius Schilder*
9899 * Various precautionary measures:
9901 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9903 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9904 (NB: This would require knowledge of the secret session ticket key
9905 to exploit, in which case you'd be SOL either way.)
9907 - Change bn_nist.c so that it will properly handle input BIGNUMs
9908 outside the expected range.
9910 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9913 *Neel Mehta, Bodo Moeller*
9915 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9916 the load fails. Useful for distros.
9918 *Ben Laurie and the FreeBSD team*
9920 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9924 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9928 * Expand ENGINE to support engine supplied SSL client certificate functions.
9930 This work was sponsored by Logica.
9934 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9935 keystores. Support for SSL/TLS client authentication too.
9936 Not compiled unless enable-capieng specified to Configure.
9938 This work was sponsored by Logica.
9942 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9943 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9944 attribute creation routines such as certificate requests and PKCS#12
9949 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9951 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9952 handshake which could lead to a client crash as found using the
9953 Codenomicon TLS test suite ([CVE-2008-1672])
9955 *Steve Henson, Mark Cox*
9957 * Fix double free in TLS server name extensions which could lead to
9958 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9962 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9964 Clear the error queue to ensure that error entries left from
9965 older function calls do not interfere with the correct operation.
9967 *Lutz Jaenicke, Erik de Castro Lopo*
9969 * Remove root CA certificates of commercial CAs:
9971 The OpenSSL project does not recommend any specific CA and does not
9972 have any policy with respect to including or excluding any CA.
9973 Therefore it does not make any sense to ship an arbitrary selection
9974 of root CA certificates with the OpenSSL software.
9978 * RSA OAEP patches to fix two separate invalid memory reads.
9979 The first one involves inputs when 'lzero' is greater than
9980 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9981 before the beginning of from). The second one involves inputs where
9982 the 'db' section contains nothing but zeroes (there is a one-byte
9983 invalid read after the end of 'db').
9985 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9987 * Partial backport from 0.9.9-dev:
9989 Introduce bn_mul_mont (dedicated Montgomery multiplication
9990 procedure) as a candidate for BIGNUM assembler implementation.
9991 While 0.9.9-dev uses assembler for various architectures, only
9992 x86_64 is available by default here in the 0.9.8 branch, and
9993 32-bit x86 is available through a compile-time setting.
9995 To try the 32-bit x86 assembler implementation, use Configure
9996 option "enable-montasm" (which exists only for this backport).
9998 As "enable-montasm" for 32-bit x86 disclaims code stability
9999 anyway, in this constellation we activate additional code
10000 backported from 0.9.9-dev for further performance improvements,
10001 namely BN_from_montgomery_word. (To enable this otherwise,
10002 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10004 *Andy Polyakov (backport partially by Bodo Moeller)*
10006 * Add TLS session ticket callback. This allows an application to set
10007 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10008 values. This is useful for key rollover for example where several key
10009 sets may exist with different names.
10013 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10014 This was broken until now in 0.9.8 releases, such that the only way
10015 a registered ENGINE could be used (assuming it initialises
10016 successfully on the host) was to explicitly set it as the default
10017 for the relevant algorithms. This is in contradiction with 0.9.7
10018 behaviour and the documentation. With this fix, when an ENGINE is
10019 registered into a given algorithm's table of implementations, the
10020 'uptodate' flag is reset so that auto-discovery will be used next
10021 time a new context for that algorithm attempts to select an
10024 *Ian Lister (tweaked by Geoff Thorpe)*
10026 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10027 implementation in the following ways:
10029 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10032 Lack of BER streaming support means one pass streaming processing is
10033 only supported if data is detached: setting the streaming flag is
10034 ignored for embedded content.
10036 CMS support is disabled by default and must be explicitly enabled
10037 with the enable-cms configuration option.
10041 * Update the GMP engine glue to do direct copies between BIGNUM and
10042 mpz_t when openssl and GMP use the same limb size. Otherwise the
10043 existing "conversion via a text string export" trick is still used.
10045 *Paul Sheer <paulsheer@gmail.com>*
10047 * Zlib compression BIO. This is a filter BIO which compressed and
10048 uncompresses any data passed through it.
10052 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10053 RFC3394 compatible AES key wrapping.
10057 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10058 sets string data without copying. X509_ALGOR_set0() and
10059 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10060 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10061 from an X509_ATTRIBUTE structure optionally checking it occurs only
10062 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10067 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10068 to get the expected BN_FLG_CONSTTIME behavior.
10070 *Bodo Moeller (Google)*
10074 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10075 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10076 - added some more tests to do_tests.pl
10077 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10078 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10079 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10080 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10081 - various changes to netware.pl to enable gcc-cross builds on Win32
10083 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10084 - various changes to fix missing prototype warnings
10085 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10086 - added AES, WHIRLPOOL and CPUID assembler code to build files
10087 - added missing AES assembler make rules to mk1mf.pl
10088 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10090 *Guenter Knauf <eflash@gmx.net>*
10092 * Implement certificate status request TLS extension defined in RFC3546.
10093 A client can set the appropriate parameters and receive the encoded
10094 OCSP response via a callback. A server can query the supplied parameters
10095 and set the encoded OCSP response in the callback. Add simplified examples
10096 to s_client and s_server.
10100 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10102 * Fix various bugs:
10103 + Binary incompatibility of ssl_ctx_st structure
10104 + DTLS interoperation with non-compliant servers
10105 + Don't call get_session_cb() without proposed session
10106 + Fix ia64 assembler code
10108 *Andy Polyakov, Steve Henson*
10110 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10112 * DTLS Handshake overhaul. There were longstanding issues with
10113 OpenSSL DTLS implementation, which were making it impossible for
10114 RFC 4347 compliant client to communicate with OpenSSL server.
10115 Unfortunately just fixing these incompatibilities would "cut off"
10116 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10117 server keeps tolerating non RFC compliant syntax. The opposite is
10118 not true, 0.9.8f client can not communicate with earlier server.
10119 This update even addresses CVE-2007-4995.
10123 * Changes to avoid need for function casts in OpenSSL: some compilers
10124 (gcc 4.2 and later) reject their use.
10125 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10128 * Add RFC4507 support to OpenSSL. This includes the corrections in
10129 RFC4507bis. The encrypted ticket format is an encrypted encoded
10130 SSL_SESSION structure, that way new session features are automatically
10133 If a client application caches session in an SSL_SESSION structure
10134 support is transparent because tickets are now stored in the encoded
10137 The SSL_CTX structure automatically generates keys for ticket
10138 protection in servers so again support should be possible
10139 with no application modification.
10141 If a client or server wishes to disable RFC4507 support then the option
10142 SSL_OP_NO_TICKET can be set.
10144 Add a TLS extension debugging callback to allow the contents of any client
10145 or server extensions to be examined.
10147 This work was sponsored by Google.
10151 * Add initial support for TLS extensions, specifically for the server_name
10152 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10153 have new members for a host name. The SSL data structure has an
10154 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10155 stored in that context to allow for session resumption, even after the
10156 SSL has been switched to a new SSL_CTX in reaction to a client's
10157 server_name extension.
10159 New functions (subject to change):
10161 SSL_get_servername()
10162 SSL_get_servername_type()
10165 New CTRL codes and macros (subject to change):
10167 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10168 - SSL_CTX_set_tlsext_servername_callback()
10169 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10170 - SSL_CTX_set_tlsext_servername_arg()
10171 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10173 openssl s_client has a new '-servername ...' option.
10175 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10176 '-key2 ...', '-servername_fatal' (subject to change). This allows
10177 testing the HostName extension for a specific single host name ('-cert'
10178 and '-key' remain fallbacks for handshakes without HostName
10179 negotiation). If the unrecognized_name alert has to be sent, this by
10180 default is a warning; it becomes fatal with the '-servername_fatal'
10183 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10185 * Add AES and SSE2 assembly language support to VC++ build.
10189 * Mitigate attack on final subtraction in Montgomery reduction.
10193 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10194 (which previously caused an internal error).
10198 * Squeeze another 10% out of IGE mode when in != out.
10202 * AES IGE mode speedup.
10204 *Dean Gaudet (Google)*
10206 * Add the Korean symmetric 128-bit cipher SEED (see
10207 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10208 add SEED ciphersuites from RFC 4162:
10210 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10211 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10212 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10213 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10215 To minimize changes between patchlevels in the OpenSSL 0.9.8
10216 series, SEED remains excluded from compilation unless OpenSSL
10217 is configured with 'enable-seed'.
10219 *KISA, Bodo Moeller*
10221 * Mitigate branch prediction attacks, which can be practical if a
10222 single processor is shared, allowing a spy process to extract
10223 information. For detailed background information, see
10224 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10225 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10226 and Necessary Software Countermeasures"). The core of the change
10227 are new versions BN_div_no_branch() and
10228 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10229 respectively, which are slower, but avoid the security-relevant
10230 conditional branches. These are automatically called by BN_div()
10231 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10232 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10233 remove a conditional branch.
10235 BN_FLG_CONSTTIME is the new name for the previous
10236 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10237 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10238 in the exponent causes BN_mod_exp_mont() to use the alternative
10239 implementation in BN_mod_exp_mont_consttime().) The old name
10240 remains as a deprecated alias.
10242 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10243 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10244 constant-time implementations for more than just exponentiation.
10245 Here too the old name is kept as a deprecated alias.
10247 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10248 the BN_BLINDING structure gets an independent copy of the
10249 modulus. This means that the previous `BIGNUM *m` argument to
10250 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10251 essentially becomes `const BIGNUM *m`, although we can't actually
10252 change this in the header file before 0.9.9. It allows
10253 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10254 enable BN_FLG_CONSTTIME.
10256 *Matthew D Wood (Intel Corp)*
10258 * In the SSL/TLS server implementation, be strict about session ID
10259 context matching (which matters if an application uses a single
10260 external cache for different purposes). Previously,
10261 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10262 set. This did ensure strict client verification, but meant that,
10263 with applications using a single external cache for quite
10264 different requirements, clients could circumvent ciphersuite
10265 restrictions for a given session ID context by starting a session
10266 in a different context.
10270 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10271 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10272 authentication-only ciphersuites.
10276 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10277 not complete and could lead to a possible single byte overflow
10278 ([CVE-2007-5135]) [Ben Laurie]
10280 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10282 * Since AES128 and AES256 (and similarly Camellia128 and
10283 Camellia256) share a single mask bit in the logic of
10284 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10285 kludge to work properly if AES128 is available and AES256 isn't
10286 (or if Camellia128 is available and Camellia256 isn't).
10290 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10291 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10292 When a point or a seed is encoded in a BIT STRING, we need to
10293 prevent the removal of trailing zero bits to get the proper DER
10294 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10295 of a NamedBitList, for which trailing 0 bits need to be removed.)
10299 * Have SSL/TLS server implementation tolerate "mismatched" record
10300 protocol version while receiving ClientHello even if the
10301 ClientHello is fragmented. (The server can't insist on the
10302 particular protocol version it has chosen before the ServerHello
10303 message has informed the client about his choice.)
10307 * Add RFC 3779 support.
10309 *Rob Austein for ARIN, Ben Laurie*
10311 * Load error codes if they are not already present instead of using a
10312 static variable. This allows them to be cleanly unloaded and reloaded.
10313 Improve header file function name parsing.
10317 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10318 or CAPABILITY handshake as required by RFCs.
10320 *Goetz Babin-Ebell*
10322 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10324 * Introduce limits to prevent malicious keys being able to
10325 cause a denial of service. ([CVE-2006-2940])
10327 *Steve Henson, Bodo Moeller*
10329 * Fix ASN.1 parsing of certain invalid structures that can result
10330 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10332 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10333 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10335 * Fix SSL client code which could crash if connecting to a
10336 malicious SSLv2 server. ([CVE-2006-4343])
10338 *Tavis Ormandy and Will Drewry, Google Security Team*
10340 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10341 match only those. Before that, "AES256-SHA" would be interpreted
10342 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10343 the same strength classification in 0.9.7h) as we currently only
10344 have a single AES bit in the ciphersuite description bitmap.
10345 That change, however, also applied to ciphersuite strings such as
10346 "RC4-MD5" that intentionally matched multiple ciphersuites --
10347 namely, SSL 2.0 ciphersuites in addition to the more common ones
10348 from SSL 3.0/TLS 1.0.
10350 So we change the selection algorithm again: Naming an explicit
10351 ciphersuite selects this one ciphersuite, and any other similar
10352 ciphersuite (same bitmap) from *other* protocol versions.
10353 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10354 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10356 Since SSL 2.0 does not have any ciphersuites for which the
10357 128/256 bit distinction would be relevant, this works for now.
10358 The proper fix will be to use different bits for AES128 and
10359 AES256, which would have avoided the problems from the beginning;
10360 however, bits are scarce, so we can only do this in a new release
10361 (not just a patchlevel) when we can change the SSL_CIPHER
10362 definition to split the single 'unsigned long mask' bitmap into
10363 multiple values to extend the available space.
10367 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10369 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10370 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10372 * Add AES IGE and biIGE modes.
10376 * Change the Unix randomness entropy gathering to use poll() when
10377 possible instead of select(), since the latter has some
10378 undesirable limitations.
10380 *Darryl Miles via Richard Levitte and Bodo Moeller*
10382 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10383 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10384 cannot be implicitly activated as part of, e.g., the "AES" alias.
10385 However, please upgrade to OpenSSL 0.9.9[-dev] for
10386 non-experimental use of the ECC ciphersuites to get TLS extension
10387 support, which is required for curve and point format negotiation
10388 to avoid potential handshake problems.
10392 * Disable rogue ciphersuites:
10394 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10395 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10396 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10398 The latter two were purportedly from
10399 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10402 Also deactivate the remaining ciphersuites from
10403 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10404 unofficial, and the ID has long expired.
10408 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10409 dual-core machines) and other potential thread-safety issues.
10413 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10414 versions), which is now available for royalty-free use
10415 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10416 Also, add Camellia TLS ciphersuites from RFC 4132.
10418 To minimize changes between patchlevels in the OpenSSL 0.9.8
10419 series, Camellia remains excluded from compilation unless OpenSSL
10420 is configured with 'enable-camellia'.
10424 * Disable the padding bug check when compression is in use. The padding
10425 bug check assumes the first packet is of even length, this is not
10426 necessarily true if compression is enabled and can result in false
10427 positives causing handshake failure. The actual bug test is ancient
10428 code so it is hoped that implementations will either have fixed it by
10429 now or any which still have the bug do not support compression.
10433 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10435 * When applying a cipher rule check to see if string match is an explicit
10436 cipher suite and only match that one cipher suite if it is.
10440 * Link in manifests for VC++ if needed.
10442 *Austin Ziegler <halostatue@gmail.com>*
10444 * Update support for ECC-based TLS ciphersuites according to
10445 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10446 TLS extensions, which are supported starting with the 0.9.9
10447 branch, not in the OpenSSL 0.9.8 branch).
10451 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10452 opaque EVP_CIPHER_CTX handling.
10456 * Fixes and enhancements to zlib compression code. We now only use
10457 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10458 to conform with the standards mentioned here:
10459 <http://www.zlib.net/DLL_FAQ.txt>
10460 Static zlib linking now works on Windows and the new --with-zlib-include
10461 --with-zlib-lib options to Configure can be used to supply the location
10462 of the headers and library. Gracefully handle case where zlib library
10467 * Several fixes and enhancements to the OID generation code. The old code
10468 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10469 handle numbers larger than ULONG_MAX, truncated printing and had a
10470 non standard OBJ_obj2txt() behaviour.
10474 * Add support for building of engines under engine/ as shared libraries
10475 under VC++ build system.
10479 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10480 Hopefully, we will not see any false combination of paths any more.
10484 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10486 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10487 (part of SSL_OP_ALL). This option used to disable the
10488 countermeasure against man-in-the-middle protocol-version
10489 rollback in the SSL 2.0 server implementation, which is a bad
10490 idea. ([CVE-2005-2969])
10492 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10493 for Information Security, National Institute of Advanced Industrial
10494 Science and Technology [AIST], Japan)*
10496 * Add two function to clear and return the verify parameter flags.
10500 * Keep cipherlists sorted in the source instead of sorting them at
10501 runtime, thus removing the need for a lock.
10505 * Avoid some small subgroup attacks in Diffie-Hellman.
10507 *Nick Mathewson and Ben Laurie*
10509 * Add functions for well-known primes.
10513 * Extended Windows CE support.
10515 *Satoshi Nakamura and Andy Polyakov*
10517 * Initialize SSL_METHOD structures at compile time instead of during
10518 runtime, thus removing the need for a lock.
10522 * Make PKCS7_decrypt() work even if no certificate is supplied by
10523 attempting to decrypt each encrypted key in turn. Add support to
10528 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10530 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10533 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10537 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10538 key into the same file any more.
10542 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10546 * Add -utf8 command line and config file option to 'ca'.
10548 *Stefan <stf@udoma.org*
10550 * Removed the macro des_crypt(), as it seems to conflict with some
10551 libraries. Use DES_crypt().
10555 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10556 involves renaming the source and generated shared-libs for
10557 both. The engines will accept the corrected or legacy ids
10558 ('ncipher' and '4758_cca' respectively) when binding. NB,
10559 this only applies when building 'shared'.
10561 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10563 * Add attribute functions to EVP_PKEY structure. Modify
10564 PKCS12_create() to recognize a CSP name attribute and
10565 use it. Make -CSP option work again in pkcs12 utility.
10569 * Add new functionality to the bn blinding code:
10570 - automatic re-creation of the BN_BLINDING parameters after
10571 a fixed number of uses (currently 32)
10572 - add new function for parameter creation
10573 - introduce flags to control the update behaviour of the
10574 BN_BLINDING parameters
10575 - hide BN_BLINDING structure
10576 Add a second BN_BLINDING slot to the RSA structure to improve
10577 performance when a single RSA object is shared among several
10582 * Add support for DTLS.
10584 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10586 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10587 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10591 * Remove buggy and incomplete DH cert support from
10592 ssl/ssl_rsa.c and ssl/s3_both.c
10596 * Use SHA-1 instead of MD5 as the default digest algorithm for
10597 the `apps/openssl` commands.
10601 * Compile clean with "-Wall -Wmissing-prototypes
10602 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10603 DEBUG_SAFESTACK must also be set.
10607 * Change ./Configure so that certain algorithms can be disabled by default.
10608 The new counterpiece to "no-xxx" is "enable-xxx".
10610 The patented RC5 and MDC2 algorithms will now be disabled unless
10611 "enable-rc5" and "enable-mdc2", respectively, are specified.
10613 (IDEA remains enabled despite being patented. This is because IDEA
10614 is frequently required for interoperability, and there is no license
10615 fee for non-commercial use. As before, "no-idea" can be used to
10616 avoid this algorithm.)
10620 * Add processing of proxy certificates (see RFC 3820). This work was
10621 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10622 EGEE (Enabling Grids for E-science in Europe).
10626 * RC4 performance overhaul on modern architectures/implementations, such
10627 as Intel P4, IA-64 and AMD64.
10631 * New utility extract-section.pl. This can be used specify an alternative
10632 section number in a pod file instead of having to treat each file as
10633 a separate case in Makefile. This can be done by adding two lines to the
10636 =for comment openssl_section:XXX
10638 The blank line is mandatory.
10642 * New arguments -certform, -keyform and -pass for s_client and s_server
10643 to allow alternative format key and certificate files and passphrase
10648 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10649 update associated structures and add various utility functions.
10651 Add new policy related verify parameters, include policy checking in
10652 standard verify code. Enhance 'smime' application with extra parameters
10653 to support policy checking and print out.
10657 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10658 Nehemiah processors. These extensions support AES encryption in hardware
10659 as well as RNG (though RNG support is currently disabled).
10661 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10663 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10667 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10669 *Andy Polyakov and a number of other people*
10671 * Improved PowerPC platform support. Most notably BIGNUM assembler
10672 implementation contributed by IBM.
10674 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10676 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10677 exponent rather than 'unsigned long'. There is a corresponding change to
10678 the new 'rsa_keygen' element of the RSA_METHOD structure.
10680 *Jelte Jansen, Geoff Thorpe*
10682 * Functionality for creating the initial serial number file is now
10683 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10685 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10686 number file to 1, which is bound to cause problems. To avoid
10687 the problems while respecting compatibility between different 0.9.7
10688 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10689 CA.pl for serial number initialization. With the new release 0.9.8,
10690 we can fix the problem directly in the 'ca' utility.)
10694 * Reduced header interdependencies by declaring more opaque objects in
10695 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10696 give fewer recursive includes, which could break lazy source code - so
10697 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10698 developers should define this symbol when building and using openssl to
10699 ensure they track the recommended behaviour, interfaces, [etc], but
10700 backwards-compatible behaviour prevails when this isn't defined.
10704 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10708 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10709 This will generate a random key of the appropriate length based on the
10710 cipher context. The EVP_CIPHER can provide its own random key generation
10711 routine to support keys of a specific form. This is used in the des and
10712 3des routines to generate a key of the correct parity. Update S/MIME
10713 code to use new functions and hence generate correct parity DES keys.
10714 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10715 valid (weak or incorrect parity).
10719 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10720 as looking them up. This is useful when the verified structure may contain
10721 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10722 present unless the new PKCS7_NO_CRL flag is asserted.
10726 * Extend ASN1 oid configuration module. It now additionally accepts the
10729 shortName = some long name, 1.2.3.4
10733 * Reimplemented the BN_CTX implementation. There is now no more static
10734 limitation on the number of variables it can handle nor the depth of the
10735 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10736 information can now expand as required, and rather than having a single
10737 static array of bignums, BN_CTX now uses a linked-list of such arrays
10738 allowing it to expand on demand whilst maintaining the usefulness of
10739 BN_CTX's "bundling".
10743 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10744 to allow all RSA operations to function using a single BN_CTX.
10748 * Preliminary support for certificate policy evaluation and checking. This
10749 is initially intended to pass the tests outlined in "Conformance Testing
10750 of Relying Party Client Certificate Path Processing Logic" v1.07.
10754 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10755 remained unused and not that useful. A variety of other little bignum
10756 tweaks and fixes have also been made continuing on from the audit (see
10761 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10762 associated ASN1, EVP and SSL functions and old ASN1 macros.
10766 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10767 and this should never fail. So the return value from the use of
10768 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10769 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10773 * BN_CTX_get() should return zero-valued bignums, providing the same
10774 initialised value as BN_new().
10776 *Geoff Thorpe, suggested by Ulf Möller*
10778 * Support for inhibitAnyPolicy certificate extension.
10782 * An audit of the BIGNUM code is underway, for which debugging code is
10783 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10784 is considered valid when processing BIGNUMs, and causes execution to
10785 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10786 further steps are taken to deliberately pollute unused data in BIGNUM
10787 structures to try and expose faulty code further on. For now, openssl will
10788 (in its default mode of operation) continue to tolerate the inconsistent
10789 forms that it has tolerated in the past, but authors and packagers should
10790 consider trying openssl and their own applications when compiled with
10791 these debugging symbols defined. It will help highlight potential bugs in
10792 their own code, and will improve the test coverage for OpenSSL itself. At
10793 some point, these tighter rules will become openssl's default to improve
10794 maintainability, though the assert()s and other overheads will remain only
10795 in debugging configurations. See bn.h for more details.
10797 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10799 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10800 that can only be obtained through BN_CTX_new() (which implicitly
10801 initialises it). The presence of this function only made it possible
10802 to overwrite an existing structure (and cause memory leaks).
10806 * Because of the callback-based approach for implementing LHASH as a
10807 template type, lh_insert() adds opaque objects to hash-tables and
10808 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10809 to clean up those corresponding objects before destroying the hash table
10810 (and losing the object pointers). So some over-zealous constifications in
10811 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10812 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10813 prototyped to have "const" restrictions on the object pointers they are
10814 given (and so aren't required to cast them away any more).
10818 * The tmdiff.h API was so ugly and minimal that our own timing utility
10819 (speed) prefers to use its own implementation. The two implementations
10820 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10821 its object type properly exposed (MS_TM) instead of casting to/from
10822 `char *`. This may still change yet if someone realises MS_TM and
10824 aren't necessarily the greatest nomenclatures - but this is what was used
10825 internally to the implementation so I've used that for now.
10829 * Ensure that deprecated functions do not get compiled when
10830 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10831 the self-tests were still using deprecated key-generation functions so
10832 these have been updated also.
10836 * Reorganise PKCS#7 code to separate the digest location functionality
10837 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10838 New function PKCS7_set_digest() to set the digest type for PKCS#7
10839 digestedData type. Add additional code to correctly generate the
10840 digestedData type and add support for this type in PKCS7 initialization
10845 * New function PKCS7_set0_type_other() this initializes a PKCS7
10846 structure of type "other".
10850 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10851 sure the loop does correctly stop and breaking ("division by zero")
10852 modulus operations are not performed. The (pre-generated) prime
10853 table crypto/bn/bn_prime.h was already correct, but it could not be
10854 re-generated on some platforms because of the "division by zero"
10855 situation in the script.
10857 *Ralf S. Engelschall*
10859 * Update support for ECC-based TLS ciphersuites according to
10860 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10861 SHA-1 now is only used for "small" curves (where the
10862 representation of a field element takes up to 24 bytes); for
10863 larger curves, the field element resulting from ECDH is directly
10864 used as premaster secret.
10866 *Douglas Stebila (Sun Microsystems Laboratories)*
10868 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10869 curve secp160r1 to the tests.
10871 *Douglas Stebila (Sun Microsystems Laboratories)*
10873 * Add the possibility to load symbols globally with DSO.
10875 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10877 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10878 control of the error stack.
10882 * Add support for STORE in ENGINE.
10886 * Add the STORE type. The intention is to provide a common interface
10887 to certificate and key stores, be they simple file-based stores, or
10888 HSM-type store, or LDAP stores, or...
10889 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10893 * Add a generic structure called OPENSSL_ITEM. This can be used to
10894 pass a list of arguments to any function as well as provide a way
10895 for a function to pass data back to the caller.
10899 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10900 works like BUF_strdup() but can be used to duplicate a portion of
10901 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10906 * Add the function sk_find_ex() which works like sk_find(), but will
10907 return an index to an element even if an exact match couldn't be
10908 found. The index is guaranteed to point at the element where the
10909 searched-for key would be inserted to preserve sorting order.
10913 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10914 takes an extra flags argument for optional functionality. Currently,
10915 the following flags are defined:
10917 OBJ_BSEARCH_VALUE_ON_NOMATCH
10918 This one gets OBJ_bsearch_ex() to return a pointer to the first
10919 element where the comparing function returns a negative or zero
10922 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10923 This one gets OBJ_bsearch_ex() to return a pointer to the first
10924 element where the comparing function returns zero. This is useful
10925 if there are more than one element where the comparing function
10930 * Make it possible to create self-signed certificates with 'openssl ca'
10931 in such a way that the self-signed certificate becomes part of the
10932 CA database and uses the same mechanisms for serial number generation
10933 as all other certificate signing. The new flag '-selfsign' enables
10934 this functionality. Adapt CA.sh and CA.pl.in.
10938 * Add functionality to check the public key of a certificate request
10939 against a given private. This is useful to check that a certificate
10940 request can be signed by that key (self-signing).
10944 * Make it possible to have multiple active certificates with the same
10945 subject in the CA index file. This is done only if the keyword
10946 'unique_subject' is set to 'no' in the main CA section (default
10947 if 'CA_default') of the configuration file. The value is saved
10948 with the database itself in a separate index attribute file,
10949 named like the index file with '.attr' appended to the name.
10953 * Generate multi-valued AVAs using '+' notation in config files for
10958 * Support for nameConstraints certificate extension.
10962 * Support for policyConstraints certificate extension.
10966 * Support for policyMappings certificate extension.
10970 * Make sure the default DSA_METHOD implementation only uses its
10971 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10972 and change its own handlers to be NULL so as to remove unnecessary
10973 indirection. This lets alternative implementations fallback to the
10974 default implementation more easily.
10978 * Support for directoryName in GeneralName related extensions
10983 * Make it possible to link applications using Makefile.shared.
10984 Make that possible even when linking against static libraries!
10988 * Support for single pass processing for S/MIME signing. This now
10989 means that S/MIME signing can be done from a pipe, in addition
10990 cleartext signing (multipart/signed type) is effectively streaming
10991 and the signed data does not need to be all held in memory.
10993 This is done with a new flag PKCS7_STREAM. When this flag is set
10994 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10995 is done after the data is output (and digests calculated) in
10996 SMIME_write_PKCS7().
11000 * Add full support for -rpath/-R, both in shared libraries and
11001 applications, at least on the platforms where it's known how
11006 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11007 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11008 will now compute a table of multiples of the generator that
11009 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11010 faster (notably in the case of a single point multiplication,
11011 scalar * generator).
11013 *Nils Larsch, Bodo Moeller*
11015 * IPv6 support for certificate extensions. The various extensions
11016 which use the IP:a.b.c.d can now take IPv6 addresses using the
11017 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11022 * Added an ENGINE that implements RSA by performing private key
11023 exponentiations with the GMP library. The conversions to and from
11024 GMP's mpz_t format aren't optimised nor are any montgomery forms
11025 cached, and on x86 it appears OpenSSL's own performance has caught up.
11026 However there are likely to be other architectures where GMP could
11027 provide a boost. This ENGINE is not built in by default, but it can be
11028 specified at Configure time and should be accompanied by the necessary
11029 linker additions, eg;
11030 ./config -DOPENSSL_USE_GMP -lgmp
11034 * "openssl engine" will not display ENGINE/DSO load failure errors when
11035 testing availability of engines with "-t" - the old behaviour is
11036 produced by increasing the feature's verbosity with "-tt".
11040 * ECDSA routines: under certain error conditions uninitialized BN objects
11041 could be freed. Solution: make sure initialization is performed early
11042 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11047 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11048 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11049 software implementations. For DSA and DH, parameter generation can
11050 also be overridden by providing the appropriate method callbacks.
11054 * Change the "progress" mechanism used in key-generation and
11055 primality testing to functions that take a new BN_GENCB pointer in
11056 place of callback/argument pairs. The new API functions have `_ex`
11057 postfixes and the older functions are reimplemented as wrappers for
11058 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11059 declarations of the old functions to help (graceful) attempts to
11060 migrate to the new functions. Also, the new key-generation API
11061 functions operate on a caller-supplied key-structure and return
11062 success/failure rather than returning a key or NULL - this is to
11063 help make "keygen" another member function of RSA_METHOD etc.
11065 Example for using the new callback interface:
11067 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11068 void *my_arg = ...;
11071 BN_GENCB_set(&my_cb, my_callback, my_arg);
11073 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11074 /* For the meaning of a, b in calls to my_callback(), see the
11075 * documentation of the function that calls the callback.
11076 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11077 * my_callback should return 1 if it wants BN_is_prime_ex()
11078 * to continue, or 0 to stop.
11083 * Change the ZLIB compression method to be stateful, and make it
11084 available to TLS with the number defined in
11085 draft-ietf-tls-compression-04.txt.
11089 * Add the ASN.1 structures and functions for CertificatePair, which
11090 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11092 CertificatePair ::= SEQUENCE {
11093 forward [0] Certificate OPTIONAL,
11094 reverse [1] Certificate OPTIONAL,
11095 -- at least one of the pair shall be present -- }
11097 Also implement the PEM functions to read and write certificate
11098 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11100 This needed to be defined, mostly for the sake of the LDAP
11101 attribute crossCertificatePair, but may prove useful elsewhere as
11106 * Make it possible to inhibit symlinking of shared libraries in
11107 Makefile.shared, for Cygwin's sake.
11111 * Extend the BIGNUM API by creating a function
11112 void BN_set_negative(BIGNUM *a, int neg);
11113 and a macro that behave like
11114 int BN_is_negative(const BIGNUM *a);
11116 to avoid the need to access 'a->neg' directly in applications.
11120 * Implement fast modular reduction for pseudo-Mersenne primes
11121 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11122 EC_GROUP_new_curve_GFp() will now automatically use this
11125 *Nils Larsch <nla@trustcenter.de>*
11127 * Add new lock type (CRYPTO_LOCK_BN).
11131 * Change the ENGINE framework to automatically load engines
11132 dynamically from specific directories unless they could be
11133 found to already be built in or loaded. Move all the
11134 current engines except for the cryptodev one to a new
11135 directory engines/.
11136 The engines in engines/ are built as shared libraries if
11137 the "shared" options was given to ./Configure or ./config.
11138 Otherwise, they are inserted in libcrypto.a.
11139 /usr/local/ssl/engines is the default directory for dynamic
11140 engines, but that can be overridden at configure time through
11141 the usual use of --prefix and/or --openssldir, and at run
11142 time with the environment variable OPENSSL_ENGINES.
11144 *Geoff Thorpe and Richard Levitte*
11146 * Add Makefile.shared, a helper makefile to build shared
11147 libraries. Adapt Makefile.org.
11151 * Add version info to Win32 DLLs.
11153 *Peter 'Luna' Runestig" <peter@runestig.com>*
11155 * Add new 'medium level' PKCS#12 API. Certificates and keys
11156 can be added using this API to created arbitrary PKCS#12
11157 files while avoiding the low-level API.
11159 New options to PKCS12_create(), key or cert can be NULL and
11160 will then be omitted from the output file. The encryption
11161 algorithm NIDs can be set to -1 for no encryption, the mac
11162 iteration count can be set to 0 to omit the mac.
11164 Enhance pkcs12 utility by making the -nokeys and -nocerts
11165 options work when creating a PKCS#12 file. New option -nomac
11166 to omit the mac, NONE can be set for an encryption algorithm.
11167 New code is modified to use the enhanced PKCS12_create()
11168 instead of the low-level API.
11172 * Extend ASN1 encoder to support indefinite length constructed
11173 encoding. This can output sequences tags and octet strings in
11174 this form. Modify pk7_asn1.c to support indefinite length
11175 encoding. This is experimental and needs additional code to
11176 be useful, such as an ASN1 bio and some enhanced streaming
11179 Extend template encode functionality so that tagging is passed
11180 down to the template encoder.
11184 * Let 'openssl req' fail if an argument to '-newkey' is not
11185 recognized instead of using RSA as a default.
11189 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11190 As these are not official, they are not included in "ALL";
11191 the "ECCdraft" ciphersuite group alias can be used to select them.
11193 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11195 * Add ECDH engine support.
11197 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11199 * Add ECDH in new directory crypto/ecdh/.
11201 *Douglas Stebila (Sun Microsystems Laboratories)*
11203 * Let BN_rand_range() abort with an error after 100 iterations
11204 without success (which indicates a broken PRNG).
11208 * Change BN_mod_sqrt() so that it verifies that the input value
11209 is really the square of the return value. (Previously,
11210 BN_mod_sqrt would show GIGO behaviour.)
11214 * Add named elliptic curves over binary fields from X9.62, SECG,
11215 and WAP/WTLS; add OIDs that were still missing.
11217 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11219 * Extend the EC library for elliptic curves over binary fields
11220 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11223 EC_GF2m_simple_method
11227 EC_GROUP_new_curve_GF2m
11228 EC_GROUP_set_curve_GF2m
11229 EC_GROUP_get_curve_GF2m
11230 EC_POINT_set_affine_coordinates_GF2m
11231 EC_POINT_get_affine_coordinates_GF2m
11232 EC_POINT_set_compressed_coordinates_GF2m
11234 Point compression for binary fields is disabled by default for
11235 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11238 As binary polynomials are represented as BIGNUMs, various members
11239 of the EC_GROUP and EC_POINT data structures can be shared
11240 between the implementations for prime fields and binary fields;
11241 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11242 are essentially identical to their `..._GFp` counterparts.
11243 (For simplicity, the `..._GFp` prefix has been dropped from
11244 various internal method names.)
11246 An internal 'field_div' method (similar to 'field_mul' and
11247 'field_sqr') has been added; this is used only for binary fields.
11249 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11251 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11252 through methods ('mul', 'precompute_mult').
11254 The generic implementations (now internally called 'ec_wNAF_mul'
11255 and 'ec_wNAF_precomputed_mult') remain the default if these
11256 methods are undefined.
11258 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11260 * New function EC_GROUP_get_degree, which is defined through
11261 EC_METHOD. For curves over prime fields, this returns the bit
11262 length of the modulus.
11264 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11266 * New functions EC_GROUP_dup, EC_POINT_dup.
11267 (These simply call ..._new and ..._copy).
11269 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11271 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11272 Polynomials are represented as BIGNUMs (where the sign bit is not
11273 used) in the following functions [macros]:
11276 BN_GF2m_sub [= BN_GF2m_add]
11277 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11278 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11279 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11281 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11282 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11283 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11284 BN_GF2m_cmp [= BN_ucmp]
11286 (Note that only the 'mod' functions are actually for fields GF(2^m).
11287 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11289 For some functions, an the irreducible polynomial defining a
11290 field can be given as an 'unsigned int[]' with strictly
11291 decreasing elements giving the indices of those bits that are set;
11292 i.e., p[] represents the polynomial
11293 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11295 p[0] > p[1] > ... > p[k] = 0.
11296 This applies to the following functions:
11299 BN_GF2m_mod_mul_arr
11300 BN_GF2m_mod_sqr_arr
11301 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11302 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11303 BN_GF2m_mod_exp_arr
11304 BN_GF2m_mod_sqrt_arr
11305 BN_GF2m_mod_solve_quad_arr
11309 Conversion can be performed by the following functions:
11314 bntest.c has additional tests for binary polynomial arithmetic.
11316 Two implementations for BN_GF2m_mod_div() are available.
11317 The default algorithm simply uses BN_GF2m_mod_inv() and
11318 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11319 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11320 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11322 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11324 * Add new error code 'ERR_R_DISABLED' that can be used when some
11325 functionality is disabled at compile-time.
11327 *Douglas Stebila <douglas.stebila@sun.com>*
11329 * Change default behaviour of 'openssl asn1parse' so that more
11330 information is visible when viewing, e.g., a certificate:
11332 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11333 mode the content of non-printable OCTET STRINGs is output in a
11334 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11335 avoid the appearance of a printable string.
11337 *Nils Larsch <nla@trustcenter.de>*
11339 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11341 EC_GROUP_set_asn1_flag()
11342 EC_GROUP_get_asn1_flag()
11343 EC_GROUP_set_point_conversion_form()
11344 EC_GROUP_get_point_conversion_form()
11345 These control ASN1 encoding details:
11346 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11347 has been set to OPENSSL_EC_NAMED_CURVE.
11348 - Points are encoded in uncompressed form by default; options for
11349 asn1_for are as for point2oct, namely
11350 POINT_CONVERSION_COMPRESSED
11351 POINT_CONVERSION_UNCOMPRESSED
11352 POINT_CONVERSION_HYBRID
11354 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11356 EC_GROUP_set_seed()
11357 EC_GROUP_get0_seed()
11358 EC_GROUP_get_seed_len()
11359 This is used only for ASN1 purposes (so far).
11361 *Nils Larsch <nla@trustcenter.de>*
11363 * Add 'field_type' member to EC_METHOD, which holds the NID
11364 of the appropriate field type OID. The new function
11365 EC_METHOD_get_field_type() returns this value.
11367 *Nils Larsch <nla@trustcenter.de>*
11370 EC_POINT_point2bn()
11371 EC_POINT_bn2point()
11372 EC_POINT_point2hex()
11373 EC_POINT_hex2point()
11374 providing useful interfaces to EC_POINT_point2oct() and
11375 EC_POINT_oct2point().
11377 *Nils Larsch <nla@trustcenter.de>*
11379 * Change internals of the EC library so that the functions
11380 EC_GROUP_set_generator()
11381 EC_GROUP_get_generator()
11382 EC_GROUP_get_order()
11383 EC_GROUP_get_cofactor()
11384 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11385 to methods, which would lead to unnecessary code duplication when
11386 adding different types of curves.
11388 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11390 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11391 arithmetic, and such that modified wNAFs are generated
11392 (which avoid length expansion in many cases).
11396 * Add a function EC_GROUP_check_discriminant() (defined via
11397 EC_METHOD) that verifies that the curve discriminant is non-zero.
11399 Add a function EC_GROUP_check() that makes some sanity tests
11400 on a EC_GROUP, its generator and order. This includes
11401 EC_GROUP_check_discriminant().
11403 *Nils Larsch <nla@trustcenter.de>*
11405 * Add ECDSA in new directory crypto/ecdsa/.
11407 Add applications 'openssl ecparam' and 'openssl ecdsa'
11408 (these are based on 'openssl dsaparam' and 'openssl dsa').
11410 ECDSA support is also included in various other files across the
11411 library. Most notably,
11412 - 'openssl req' now has a '-newkey ecdsa:file' option;
11413 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11414 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11415 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11416 them suitable for ECDSA where domain parameters must be
11417 extracted before the specific public key;
11418 - ECDSA engine support has been added.
11420 *Nils Larsch <nla@trustcenter.de>*
11422 * Include some named elliptic curves, and add OIDs from X9.62,
11423 SECG, and WAP/WTLS. Each curve can be obtained from the new
11425 EC_GROUP_new_by_curve_name(),
11426 and the list of available named curves can be obtained with
11427 EC_get_builtin_curves().
11428 Also add a 'curve_name' member to EC_GROUP objects, which can be
11430 EC_GROUP_set_curve_name()
11431 EC_GROUP_get_curve_name()
11433 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11435 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11436 was actually never needed) and in BN_mul(). The removal in BN_mul()
11437 required a small change in bn_mul_part_recursive() and the addition
11438 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11439 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11440 bn_sub_words() and bn_add_words() except they take arrays with
11445 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11447 * Cleanse PEM buffers before freeing them since they may contain
11450 *Benjamin Bennett <ben@psc.edu>*
11452 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11453 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11454 authentication-only ciphersuites.
11458 * Since AES128 and AES256 share a single mask bit in the logic of
11459 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11460 kludge to work properly if AES128 is available and AES256 isn't.
11464 * Expand security boundary to match 1.1.1 module.
11468 * Remove redundant features: hash file source, editing of test vectors
11469 modify fipsld to use external fips_premain.c signature.
11473 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11474 run algorithm test programs.
11478 * Make algorithm test programs more tolerant of whitespace.
11482 * Have SSL/TLS server implementation tolerate "mismatched" record
11483 protocol version while receiving ClientHello even if the
11484 ClientHello is fragmented. (The server can't insist on the
11485 particular protocol version it has chosen before the ServerHello
11486 message has informed the client about his choice.)
11490 * Load error codes if they are not already present instead of using a
11491 static variable. This allows them to be cleanly unloaded and reloaded.
11495 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11497 * Introduce limits to prevent malicious keys being able to
11498 cause a denial of service. ([CVE-2006-2940])
11500 *Steve Henson, Bodo Moeller*
11502 * Fix ASN.1 parsing of certain invalid structures that can result
11503 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11505 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11506 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11508 * Fix SSL client code which could crash if connecting to a
11509 malicious SSLv2 server. ([CVE-2006-4343])
11511 *Tavis Ormandy and Will Drewry, Google Security Team*
11513 * Change ciphersuite string processing so that an explicit
11514 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11515 will no longer include "AES128-SHA"), and any other similar
11516 ciphersuite (same bitmap) from *other* protocol versions (so that
11517 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11518 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11519 changes from 0.9.8b and 0.9.8d.
11523 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11525 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11526 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11528 * Change the Unix randomness entropy gathering to use poll() when
11529 possible instead of select(), since the latter has some
11530 undesirable limitations.
11532 *Darryl Miles via Richard Levitte and Bodo Moeller*
11534 * Disable rogue ciphersuites:
11536 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11537 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11538 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11540 The latter two were purportedly from
11541 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11544 Also deactivate the remaining ciphersuites from
11545 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11546 unofficial, and the ID has long expired.
11550 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11551 dual-core machines) and other potential thread-safety issues.
11555 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11557 * Adapt fipsld and the build system to link against the validated FIPS
11558 module in FIPS mode.
11562 * Fixes for VC++ 2005 build under Windows.
11566 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11567 from a Windows bash shell such as MSYS. It is autodetected from the
11568 "config" script when run from a VC++ environment. Modify standard VC++
11569 build to use fipscanister.o from the GNU make build.
11573 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11575 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11576 The value now differs depending on if you build for FIPS or not.
11577 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11578 safely run with a non-FIPSed libcrypto, as it may crash because of
11579 the difference induced by this change.
11583 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11585 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11586 (part of SSL_OP_ALL). This option used to disable the
11587 countermeasure against man-in-the-middle protocol-version
11588 rollback in the SSL 2.0 server implementation, which is a bad
11589 idea. ([CVE-2005-2969])
11591 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11592 for Information Security, National Institute of Advanced Industrial
11593 Science and Technology [AIST, Japan)]*
11595 * Minimal support for X9.31 signatures and PSS padding modes. This is
11596 mainly for FIPS compliance and not fully integrated at this stage.
11600 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11601 the exponentiation using a fixed-length exponent. (Otherwise,
11602 the information leaked through timing could expose the secret key
11603 after many signatures; cf. Bleichenbacher's attack on DSA with
11608 * Make a new fixed-window mod_exp implementation the default for
11609 RSA, DSA, and DH private-key operations so that the sequence of
11610 squares and multiplies and the memory access pattern are
11611 independent of the particular secret key. This will mitigate
11612 cache-timing and potential related attacks.
11614 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11615 and this is automatically used by BN_mod_exp_mont() if the new flag
11616 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11617 will use this BN flag for private exponents unless the flag
11618 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11619 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11621 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11623 * Change the client implementation for SSLv23_method() and
11624 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11625 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11626 (Previously, the SSL 2.0 backwards compatible Client Hello
11627 message format would be used even with SSL_OP_NO_SSLv2.)
11631 * Add support for smime-type MIME parameter in S/MIME messages which some
11636 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11637 a threadsafe manner. Modify rsa code to use new function and add calls
11638 to dsa and dh code (which had race conditions before).
11642 * Include the fixed error library code in the C error file definitions
11643 instead of fixing them up at runtime. This keeps the error code
11644 structures constant.
11648 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11650 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11653 * Fixes for newer kerberos headers. NB: the casts are needed because
11654 the 'length' field is signed on one version and unsigned on another
11655 with no (?) obvious way to tell the difference, without these VC++
11656 complains. Also the "definition" of FAR (blank) is no longer included
11657 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11658 some needed definitions.
11662 * Undo Cygwin change.
11666 * Added support for proxy certificates according to RFC 3820.
11667 Because they may be a security thread to unaware applications,
11668 they must be explicitly allowed in run-time. See
11669 docs/HOWTO/proxy_certificates.txt for further information.
11673 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11675 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11676 server and client random values. Previously
11677 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11678 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11680 This change has negligible security impact because:
11682 1. Server and client random values still have 24 bytes of pseudo random
11685 2. Server and client random values are sent in the clear in the initial
11688 3. The master secret is derived using the premaster secret (48 bytes in
11689 size for static RSA ciphersuites) as well as client server and random
11692 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11695 *Stephen Henson, reported by UK NISCC*
11697 * Use Windows randomness collection on Cygwin.
11701 * Fix hang in EGD/PRNGD query when communication socket is closed
11702 prematurely by EGD/PRNGD.
11704 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11706 * Prompt for pass phrases when appropriate for PKCS12 input format.
11710 * Back-port of selected performance improvements from development
11711 branch, as well as improved support for PowerPC platforms.
11715 * Add lots of checks for memory allocation failure, error codes to indicate
11716 failure and freeing up memory if a failure occurs.
11718 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11720 * Add new -passin argument to dgst.
11724 * Perform some character comparisons of different types in X509_NAME_cmp:
11725 this is needed for some certificates that re-encode DNs into UTF8Strings
11726 (in violation of RFC3280) and can't or won't issue name rollover
11731 * Make an explicit check during certificate validation to see that
11732 the CA setting in each certificate on the chain is correct. As a
11733 side effect always do the following basic checks on extensions,
11734 not just when there's an associated purpose to the check:
11736 - if there is an unhandled critical extension (unless the user
11737 has chosen to ignore this fault)
11738 - if the path length has been exceeded (if one is set at all)
11739 - that certain extensions fit the associated purpose (if one has
11744 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11746 * Avoid a race condition when CRLs are checked in a multi threaded
11747 environment. This would happen due to the reordering of the revoked
11748 entries during signature checking and serial number lookup. Now the
11749 encoding is cached and the serial number sort performed under a lock.
11750 Add new STACK function sk_is_sorted().
11754 * Add Delta CRL to the extension code.
11758 * Various fixes to s3_pkt.c so alerts are sent properly.
11760 *David Holmes <d.holmes@f5.com>*
11762 * Reduce the chances of duplicate issuer name and serial numbers (in
11763 violation of RFC3280) using the OpenSSL certificate creation utilities.
11764 This is done by creating a random 64 bit value for the initial serial
11765 number when a serial number file is created or when a self signed
11766 certificate is created using 'openssl req -x509'. The initial serial
11767 number file is created using 'openssl x509 -next_serial' in CA.pl
11768 rather than being initialized to 1.
11772 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11774 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11775 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11777 *Joe Orton, Steve Henson*
11779 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11782 *Joe Orton, Steve Henson*
11784 * Make it possible to have multiple active certificates with the same
11785 subject in the CA index file. This is done only if the keyword
11786 'unique_subject' is set to 'no' in the main CA section (default
11787 if 'CA_default') of the configuration file. The value is saved
11788 with the database itself in a separate index attribute file,
11789 named like the index file with '.attr' appended to the name.
11793 * X509 verify fixes. Disable broken certificate workarounds when
11794 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11795 keyUsage extension present. Don't accept CRLs with unhandled critical
11796 extensions: since verify currently doesn't process CRL extensions this
11797 rejects a CRL with *any* critical extensions. Add new verify error codes
11802 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11803 A clarification of RFC2560 will require the use of OCTET STRINGs and
11804 some implementations cannot handle the current raw format. Since OpenSSL
11805 copies and compares OCSP nonces as opaque blobs without any attempt at
11806 parsing them this should not create any compatibility issues.
11810 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11811 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11812 this HMAC (and other) operations are several times slower than OpenSSL
11817 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11819 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11821 * Use the correct content when signing type "other".
11825 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11827 * Fix various bugs revealed by running the NISCC test suite:
11829 Stop out of bounds reads in the ASN1 code when presented with
11830 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11832 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11834 If verify callback ignores invalid public key errors don't try to check
11835 certificate signature with the NULL public key.
11839 * New -ignore_err option in ocsp application to stop the server
11840 exiting on the first error in a request.
11844 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11845 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11850 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11851 extra data after the compression methods not only for TLS 1.0
11852 but also for SSL 3.0 (as required by the specification).
11854 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11856 * Change X509_certificate_type() to mark the key as exported/exportable
11857 when it's 512 *bits* long, not 512 bytes.
11861 * Change AES_cbc_encrypt() so it outputs exact multiple of
11862 blocks during encryption.
11866 * Various fixes to base64 BIO and non blocking I/O. On write
11867 flushes were not handled properly if the BIO retried. On read
11868 data was not being buffered properly and had various logic bugs.
11869 This also affects blocking I/O when the data being decoded is a
11874 * Various S/MIME bugfixes and compatibility changes:
11875 output correct application/pkcs7 MIME type if
11876 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11877 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11878 of files as .eml work). Correctly handle very long lines in MIME
11883 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11885 * Countermeasure against the Klima-Pokorny-Rosa extension of
11886 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11887 a protocol version number mismatch like a decryption error
11888 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11892 * Turn on RSA blinding by default in the default implementation
11893 to avoid a timing attack. Applications that don't want it can call
11894 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11895 They would be ill-advised to do so in most cases.
11897 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11899 * Change RSA blinding code so that it works when the PRNG is not
11900 seeded (in this case, the secret RSA exponent is abused as
11901 an unpredictable seed -- if it is not unpredictable, there
11902 is no point in blinding anyway). Make RSA blinding thread-safe
11903 by remembering the creator's thread ID in rsa->blinding and
11904 having all other threads use local one-time blinding factors
11905 (this requires more computation than sharing rsa->blinding, but
11906 avoids excessive locking; and if an RSA object is not shared
11907 between threads, blinding will still be very fast).
11911 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11912 ENGINE as defaults for all supported algorithms irrespective of
11913 the 'flags' parameter. 'flags' is now honoured, so applications
11914 should make sure they are passing it correctly.
11918 * Target "mingw" now allows native Windows code to be generated in
11919 the Cygwin environment as well as with the MinGW compiler.
11923 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11925 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11926 via timing by performing a MAC computation even if incorrect
11927 block cipher padding has been found. This is a countermeasure
11928 against active attacks where the attacker has to distinguish
11929 between bad padding and a MAC verification error. ([CVE-2003-0078])
11931 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11932 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11933 Martin Vuagnoux (EPFL, Ilion)*
11935 * Make the no-err option work as intended. The intention with no-err
11936 is not to have the whole error stack handling routines removed from
11937 libcrypto, it's only intended to remove all the function name and
11938 reason texts, thereby removing some of the footprint that may not
11939 be interesting if those errors aren't displayed anyway.
11941 NOTE: it's still possible for any application or module to have its
11942 own set of error texts inserted. The routines are there, just not
11943 used by default when no-err is given.
11947 * Add support for FreeBSD on IA64.
11949 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11951 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11952 Kerberos function mit_des_cbc_cksum(). Before this change,
11953 the value returned by DES_cbc_cksum() was like the one from
11954 mit_des_cbc_cksum(), except the bytes were swapped.
11956 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11958 * Allow an application to disable the automatic SSL chain building.
11959 Before this a rather primitive chain build was always performed in
11960 ssl3_output_cert_chain(): an application had no way to send the
11961 correct chain if the automatic operation produced an incorrect result.
11963 Now the chain builder is disabled if either:
11965 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11967 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11969 The reasoning behind this is that an application would not want the
11970 auto chain building to take place if extra chain certificates are
11971 present and it might also want a means of sending no additional
11972 certificates (for example the chain has two certificates and the
11977 * Add the possibility to build without the ENGINE framework.
11979 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11981 * Under Win32 gmtime() can return NULL: check return value in
11982 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11986 * DSA routines: under certain error conditions uninitialized BN objects
11987 could be freed. Solution: make sure initialization is performed early
11988 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11989 Nils Larsch <nla@trustcenter.de> via PR#459)
11993 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11994 checked on reconnect on the client side, therefore session resumption
11995 could still fail with a "ssl session id is different" error. This
11996 behaviour is masked when SSL_OP_ALL is used due to
11997 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11998 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11999 followup to PR #377.
12003 * IA-32 assembler support enhancements: unified ELF targets, support
12004 for SCO/Caldera platforms, fix for Cygwin shared build.
12008 * Add support for FreeBSD on sparc64. As a consequence, support for
12009 FreeBSD on non-x86 processors is separate from x86 processors on
12010 the config script, much like the NetBSD support.
12012 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12014 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12016 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12019 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12020 code (06) was taken as the first octet of the session ID and the last
12021 octet was ignored consequently. As a result SSLv2 client side session
12022 caching could not have worked due to the session ID mismatch between
12024 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12029 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12030 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12035 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12036 seems that in spite of existing for more than a year, many application
12037 author have done nothing to provide the necessary callbacks, which
12038 means that this particular engine will not work properly anywhere.
12039 This is a very unfortunate situation which forces us, in the name
12040 of usability, to give the hw_ncipher.c a static lock, which is part
12042 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12043 appear in 0.9.8 or later. We EXPECT application authors to have
12044 dealt properly with this when 0.9.8 is released (unless we actually
12045 make such changes in the libcrypto locking code that changes will
12046 have to be made anyway).
12050 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12051 octets have been read, EOF or an error occurs. Without this change
12052 some truncated ASN1 structures will not produce an error.
12056 * Disable Heimdal support, since it hasn't been fully implemented.
12057 Still give the possibility to force the use of Heimdal, but with
12058 warnings and a request that patches get sent to openssl-dev.
12062 * Add the VC-CE target, introduce the WINCE sysname, and add
12063 INSTALL.WCE and appropriate conditionals to make it build.
12065 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12067 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12068 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12069 edit numbers of the version.
12071 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12073 * Introduce safe string copy and catenation functions
12074 (BUF_strlcpy() and BUF_strlcat()).
12076 *Ben Laurie (CHATS) and Richard Levitte*
12078 * Avoid using fixed-size buffers for one-line DNs.
12080 *Ben Laurie (CHATS)*
12082 * Add BUF_MEM_grow_clean() to avoid information leakage when
12083 resizing buffers containing secrets, and use where appropriate.
12085 *Ben Laurie (CHATS)*
12087 * Avoid using fixed size buffers for configuration file location.
12089 *Ben Laurie (CHATS)*
12091 * Avoid filename truncation for various CA files.
12093 *Ben Laurie (CHATS)*
12095 * Use sizeof in preference to magic numbers.
12097 *Ben Laurie (CHATS)*
12099 * Avoid filename truncation in cert requests.
12101 *Ben Laurie (CHATS)*
12103 * Add assertions to check for (supposedly impossible) buffer
12106 *Ben Laurie (CHATS)*
12108 * Don't cache truncated DNS entries in the local cache (this could
12109 potentially lead to a spoofing attack).
12111 *Ben Laurie (CHATS)*
12113 * Fix various buffers to be large enough for hex/decimal
12114 representations in a platform independent manner.
12116 *Ben Laurie (CHATS)*
12118 * Add CRYPTO_realloc_clean() to avoid information leakage when
12119 resizing buffers containing secrets, and use where appropriate.
12121 *Ben Laurie (CHATS)*
12123 * Add BIO_indent() to avoid much slightly worrying code to do
12126 *Ben Laurie (CHATS)*
12128 * Convert sprintf()/BIO_puts() to BIO_printf().
12130 *Ben Laurie (CHATS)*
12132 * buffer_gets() could terminate with the buffer only half
12135 *Ben Laurie (CHATS)*
12137 * Add assertions to prevent user-supplied crypto functions from
12138 overflowing internal buffers by having large block sizes, etc.
12140 *Ben Laurie (CHATS)*
12142 * New OPENSSL_assert() macro (similar to assert(), but enabled
12145 *Ben Laurie (CHATS)*
12147 * Eliminate unused copy of key in RC4.
12149 *Ben Laurie (CHATS)*
12151 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12153 *Ben Laurie (CHATS)*
12155 * Fix off-by-one error in EGD path.
12157 *Ben Laurie (CHATS)*
12159 * If RANDFILE path is too long, ignore instead of truncating.
12161 *Ben Laurie (CHATS)*
12163 * Eliminate unused and incorrectly sized X.509 structure
12166 *Ben Laurie (CHATS)*
12168 * Eliminate unused and dangerous function knumber().
12170 *Ben Laurie (CHATS)*
12172 * Eliminate unused and dangerous structure, KSSL_ERR.
12174 *Ben Laurie (CHATS)*
12176 * Protect against overlong session ID context length in an encoded
12177 session object. Since these are local, this does not appear to be
12180 *Ben Laurie (CHATS)*
12182 * Change from security patch (see 0.9.6e below) that did not affect
12183 the 0.9.6 release series:
12185 Remote buffer overflow in SSL3 protocol - an attacker could
12186 supply an oversized master key in Kerberos-enabled versions.
12189 *Ben Laurie (CHATS)*
12191 * Change the SSL kerb5 codes to match RFC 2712.
12195 * Make -nameopt work fully for req and add -reqopt switch.
12197 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12199 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12201 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12203 * Make sure tests can be performed even if the corresponding algorithms
12204 have been removed entirely. This was also the last step to make
12205 OpenSSL compilable with DJGPP under all reasonable conditions.
12207 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12209 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12210 to allow version independent disabling of normally unselected ciphers,
12211 which may be activated as a side-effect of selecting a single cipher.
12213 (E.g., cipher list string "RSA" enables ciphersuites that are left
12214 out of "ALL" because they do not provide symmetric encryption.
12215 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12217 *Lutz Jaenicke, Bodo Moeller*
12219 * Add appropriate support for separate platform-dependent build
12220 directories. The recommended way to make a platform-dependent
12221 build directory is the following (tested on Linux), maybe with
12224 # Place yourself outside of the OpenSSL source tree. In
12225 # this example, the environment variable OPENSSL_SOURCE
12226 # is assumed to contain the absolute OpenSSL source directory.
12227 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12228 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12229 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12230 mkdir -p `dirname $F`
12231 ln -s $OPENSSL_SOURCE/$F $F
12234 To be absolutely sure not to disturb the source tree, a "make clean"
12235 is a good thing. If it isn't successful, don't worry about it,
12236 it probably means the source directory is very clean.
12240 * Make sure any ENGINE control commands make local copies of string
12241 pointers passed to them whenever necessary. Otherwise it is possible
12242 the caller may have overwritten (or deallocated) the original string
12243 data when a later ENGINE operation tries to use the stored values.
12245 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12247 * Improve diagnostics in file reading and command-line digests.
12249 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12251 * Add AES modes CFB and OFB to the object database. Correct an
12252 error in AES-CFB decryption.
12256 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12257 allows existing EVP_CIPHER_CTX structures to be reused after
12258 calling `EVP_*Final()`. This behaviour is used by encryption
12259 BIOs and some applications. This has the side effect that
12260 applications must explicitly clean up cipher contexts with
12261 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12265 * Check the values of dna and dnb in bn_mul_recursive before calling
12266 bn_mul_comba (a non zero value means the a or b arrays do not contain
12267 n2 elements) and fallback to bn_mul_normal if either is not zero.
12271 * Fix escaping of non-ASCII characters when using the -subj option
12272 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12276 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12277 form for "surname", serialNumber has no short form.
12278 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12279 therefore remove "mail" short name for "internet 7".
12280 The OID for unique identifiers in X509 certificates is
12281 x500UniqueIdentifier, not uniqueIdentifier.
12282 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12286 * Add an "init" command to the ENGINE config module and auto initialize
12287 ENGINEs. Without any "init" command the ENGINE will be initialized
12288 after all ctrl commands have been executed on it. If init=1 the
12289 ENGINE is initialized at that point (ctrls before that point are run
12290 on the uninitialized ENGINE and after on the initialized one). If
12291 init=0 then the ENGINE will not be initialized at all.
12295 * Fix the 'app_verify_callback' interface so that the user-defined
12296 argument is actually passed to the callback: In the
12297 SSL_CTX_set_cert_verify_callback() prototype, the callback
12298 declaration has been changed from
12301 int (*cb)(X509_STORE_CTX *,void *);
12302 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12303 i=s->ctx->app_verify_callback(&ctx)
12304 has been changed into
12305 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12307 To update applications using SSL_CTX_set_cert_verify_callback(),
12308 a dummy argument can be added to their callback functions.
12310 *D. K. Smetters <smetters@parc.xerox.com>*
12312 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12314 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12316 * Add and OPENSSL_LOAD_CONF define which will cause
12317 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12318 This allows older applications to transparently support certain
12319 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12320 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12321 load the config file and OPENSSL_add_all_algorithms_conf() which will
12322 always load it have also been added.
12326 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12327 Adjust NIDs and EVP layer.
12329 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12331 * Config modules support in openssl utility.
12333 Most commands now load modules from the config file,
12334 though in a few (such as version) this isn't done
12335 because it couldn't be used for anything.
12337 In the case of ca and req the config file used is
12338 the same as the utility itself: that is the -config
12339 command line option can be used to specify an
12344 * Move default behaviour from OPENSSL_config(). If appname is NULL
12345 use "openssl_conf" if filename is NULL use default openssl config file.
12349 * Add an argument to OPENSSL_config() to allow the use of an alternative
12350 config section name. Add a new flag to tolerate a missing config file
12351 and move code to CONF_modules_load_file().
12355 * Support for crypto accelerator cards from Accelerated Encryption
12356 Processing, www.aep.ie. (Use engine 'aep')
12357 The support was copied from 0.9.6c [engine] and adapted/corrected
12358 to work with the new engine framework.
12360 *AEP Inc. and Richard Levitte*
12362 * Support for SureWare crypto accelerator cards from Baltimore
12363 Technologies. (Use engine 'sureware')
12364 The support was copied from 0.9.6c [engine] and adapted
12365 to work with the new engine framework.
12369 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12370 make the newer ENGINE framework commands for the CHIL engine work.
12372 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12374 * Make it possible to produce shared libraries on ReliantUNIX.
12376 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12378 * Add the configuration target debug-linux-ppro.
12379 Make 'openssl rsa' use the general key loading routines
12380 implemented in `apps.c`, and make those routines able to
12381 handle the key format FORMAT_NETSCAPE and the variant
12384 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12386 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12388 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12390 * Add -keyform to rsautl, and document -engine.
12392 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12394 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12395 BIO_R_NO_SUCH_FILE error code rather than the generic
12396 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12400 * Add new functions
12401 ERR_peek_last_error
12402 ERR_peek_last_error_line
12403 ERR_peek_last_error_line_data.
12404 These are similar to
12406 ERR_peek_error_line
12407 ERR_peek_error_line_data,
12408 but report on the latest error recorded rather than the first one
12409 still in the error queue.
12411 *Ben Laurie, Bodo Moeller*
12413 * default_algorithms option in ENGINE config module. This allows things
12415 default_algorithms = ALL
12416 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12420 * Preliminary ENGINE config module.
12424 * New experimental application configuration code.
12428 * Change the AES code to follow the same name structure as all other
12429 symmetric ciphers, and behave the same way. Move everything to
12430 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12432 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12434 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12436 *Ben Laurie and Theo de Raadt*
12438 * Add option to output public keys in req command.
12440 *Massimiliano Pala madwolf@openca.org*
12442 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12443 (up to about 10% better than before for P-192 and P-224).
12447 * New functions/macros
12449 SSL_CTX_set_msg_callback(ctx, cb)
12450 SSL_CTX_set_msg_callback_arg(ctx, arg)
12451 SSL_set_msg_callback(ssl, cb)
12452 SSL_set_msg_callback_arg(ssl, arg)
12454 to request calling a callback function
12456 void cb(int write_p, int version, int content_type,
12457 const void *buf, size_t len, SSL *ssl, void *arg)
12459 whenever a protocol message has been completely received
12460 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12461 protocol version according to which the SSL library interprets
12462 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12463 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12464 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12465 specification (change_cipher_spec(20), alert(21), handshake(22)).
12466 'buf' and 'len' point to the actual message, 'ssl' to the
12467 SSL object, and 'arg' is the application-defined value set by
12468 SSL[_CTX]_set_msg_callback_arg().
12470 'openssl s_client' and 'openssl s_server' have new '-msg' options
12471 to enable a callback that displays all protocol messages.
12475 * Change the shared library support so shared libraries are built as
12476 soon as the corresponding static library is finished, and thereby get
12477 openssl and the test programs linked against the shared library.
12478 This still only happens when the keyword "shard" has been given to
12479 the configuration scripts.
12481 NOTE: shared library support is still an experimental thing, and
12482 backward binary compatibility is still not guaranteed.
12484 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12486 * Add support for Subject Information Access extension.
12488 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12490 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12491 additional bytes when new memory had to be allocated, not just
12492 when reusing an existing buffer.
12496 * New command line and configuration option 'utf8' for the req command.
12497 This allows field values to be specified as UTF8 strings.
12501 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12502 runs for the former and machine-readable output for the latter.
12506 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12507 of the e-mail address in the DN (i.e., it will go into a certificate
12508 extension only). The new configuration file option 'email_in_dn = no'
12509 has the same effect.
12511 *Massimiliano Pala madwolf@openca.org*
12513 * Change all functions with names starting with `des_` to be starting
12514 with `DES_` instead. Add wrappers that are compatible with libdes,
12515 but are named `_ossl_old_des_*`. Finally, add macros that map the
12516 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12517 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12518 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12521 Since we provide two compatibility mappings, the user needs to
12522 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12523 compatibility is desired. The default (i.e., when that macro
12524 isn't defined) is OpenSSL 0.9.6c compatibility.
12526 There are also macros that enable and disable the support of old
12527 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12528 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12529 are defined, the default will apply: to support the old des routines.
12531 In either case, one must include openssl/des.h to get the correct
12532 definitions. Do not try to just include openssl/des_old.h, that
12535 NOTE: This is a major break of an old API into a new one. Software
12536 authors are encouraged to switch to the `DES_` style functions. Some
12537 time in the future, des_old.h and the libdes compatibility functions
12538 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12539 default), and then completely removed.
12543 * Test for certificates which contain unsupported critical extensions.
12544 If such a certificate is found during a verify operation it is
12545 rejected by default: this behaviour can be overridden by either
12546 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12547 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12548 X509_supported_extension() has also been added which returns 1 if a
12549 particular extension is supported.
12553 * Modify the behaviour of EVP cipher functions in similar way to digests
12554 to retain compatibility with existing code.
12558 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12559 compatibility with existing code. In particular the 'ctx' parameter does
12560 not have to be to be initialized before the call to EVP_DigestInit() and
12561 it is tidied up after a call to EVP_DigestFinal(). New function
12562 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12563 EVP_MD_CTX_copy() changed to not require the destination to be
12564 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12565 requires the destination to be valid.
12567 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12568 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12572 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12573 so that complete 'Handshake' protocol structures are kept in memory
12574 instead of overwriting 'msg_type' and 'length' with 'body' data.
12578 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12580 *Massimo Santin via Richard Levitte*
12582 * Major restructuring to the underlying ENGINE code. This includes
12583 reduction of linker bloat, separation of pure "ENGINE" manipulation
12584 (initialisation, etc) from functionality dealing with implementations
12585 of specific crypto interfaces. This change also introduces integrated
12586 support for symmetric ciphers and digest implementations - so ENGINEs
12587 can now accelerate these by providing EVP_CIPHER and EVP_MD
12588 implementations of their own. This is detailed in
12589 [crypto/engine/README.md](crypto/engine/README.md)
12590 as it couldn't be adequately described here. However, there are a few
12591 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12592 were changed in the original introduction of ENGINE code have now
12593 reverted back - the hooking from this code to ENGINE is now a good
12594 deal more passive and at run-time, operations deal directly with
12595 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12596 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12597 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12598 they were not being used by the framework as there is no concept of a
12599 BIGNUM_METHOD and they could not be generalised to the new
12600 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12601 ENGINE_cpy() has been removed as it cannot be consistently defined in
12606 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12610 * Change mkdef.pl to sort symbols that get the same entry number,
12611 and make sure the automatically generated functions `ERR_load_*`
12612 become part of libeay.num as well.
12616 * New function SSL_renegotiate_pending(). This returns true once
12617 renegotiation has been requested (either SSL_renegotiate() call
12618 or HelloRequest/ClientHello received from the peer) and becomes
12619 false once a handshake has been completed.
12620 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12621 sends a HelloRequest, but does not ensure that a handshake takes
12622 place. SSL_renegotiate_pending() is useful for checking if the
12623 client has followed the request.)
12627 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12628 By default, clients may request session resumption even during
12629 renegotiation (if session ID contexts permit); with this option,
12630 session resumption is possible only in the first handshake.
12632 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12633 more bits available for options that should not be part of
12634 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12638 * Add some demos for certificate and certificate request creation.
12642 * Make maximum certificate chain size accepted from the peer application
12643 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12644 "Douglas E. Engert" <deengert@anl.gov>.
12648 * Add support for shared libraries for Unixware-7
12649 (Boyd Lynn Gerber <gerberb@zenez.com>).
12653 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12654 be done prior to destruction. Use this to unload error strings from
12655 ENGINEs that load their own error strings. NB: This adds two new API
12656 functions to "get" and "set" this destroy handler in an ENGINE.
12660 * Alter all existing ENGINE implementations (except "openssl" and
12661 "openbsd") to dynamically instantiate their own error strings. This
12662 makes them more flexible to be built both as statically-linked ENGINEs
12663 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12664 Also, add stub code to each that makes building them as self-contained
12665 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12669 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12670 implementations into applications that are completely implemented in
12671 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12672 commands that can be used to configure what shared-library to load and
12673 to control aspects of the way it is handled. Also, made an update to
12674 the [README-Engine.md](README-Engine.md) file
12675 that brings its information up-to-date and
12676 provides some information and instructions on the "dynamic" ENGINE
12677 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12681 * Make it possible to unload ranges of ERR strings with a new
12682 "ERR_unload_strings" function.
12686 * Add a copy() function to EVP_MD.
12690 * Make EVP_MD routines take a context pointer instead of just the
12691 md_data void pointer.
12695 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12696 that the digest can only process a single chunk of data
12697 (typically because it is provided by a piece of
12698 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12699 is only going to provide a single chunk of data, and hence the
12700 framework needn't accumulate the data for oneshot drivers.
12704 * As with "ERR", make it possible to replace the underlying "ex_data"
12705 functions. This change also alters the storage and management of global
12706 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12707 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12708 index counters. The API functions that use this state have been changed
12709 to take a "class_index" rather than pointers to the class's local STACK
12710 and counter, and there is now an API function to dynamically create new
12711 classes. This centralisation allows us to (a) plug a lot of the
12712 thread-safety problems that existed, and (b) makes it possible to clean
12713 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12714 such data would previously have always leaked in application code and
12715 workarounds were in place to make the memory debugging turn a blind eye
12716 to it. Application code that doesn't use this new function will still
12717 leak as before, but their memory debugging output will announce it now
12718 rather than letting it slide.
12720 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12721 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12722 has a return value to indicate success or failure.
12726 * Make it possible to replace the underlying "ERR" functions such that the
12727 global state (2 LHASH tables and 2 locks) is only used by the "default"
12728 implementation. This change also adds two functions to "get" and "set"
12729 the implementation prior to it being automatically set the first time
12730 any other ERR function takes place. Ie. an application can call "get",
12731 pass the return value to a module it has just loaded, and that module
12732 can call its own "set" function using that value. This means the
12733 module's "ERR" operations will use (and modify) the error state in the
12734 application and not in its own statically linked copy of OpenSSL code.
12738 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12739 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12740 the operation, and provides a more encapsulated way for external code
12741 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12742 to use these functions rather than manually incrementing the counts.
12744 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12748 * Add EVP test program.
12752 * Add symmetric cipher support to ENGINE. Expect the API to change!
12756 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12757 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12758 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12759 These allow a CRL to be built without having to access X509_CRL fields
12760 directly. Modify 'ca' application to use new functions.
12764 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12765 bug workarounds. Rollback attack detection is a security feature.
12766 The problem will only arise on OpenSSL servers when TLSv1 is not
12767 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12768 Software authors not wanting to support TLSv1 will have special reasons
12769 for their choice and can explicitly enable this option.
12771 *Bodo Moeller, Lutz Jaenicke*
12773 * Rationalise EVP so it can be extended: don't include a union of
12774 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12775 (similar to those existing for EVP_CIPHER_CTX).
12780 EVP_MD_CTX_init(&md); /* new function call */
12781 EVP_DigestInit(&md, EVP_sha1());
12782 EVP_DigestUpdate(&md, in, len);
12783 EVP_DigestFinal(&md, out, NULL);
12784 EVP_MD_CTX_cleanup(&md); /* new function call */
12788 * Make DES key schedule conform to the usual scheme, as well as
12789 correcting its structure. This means that calls to DES functions
12790 now have to pass a pointer to a des_key_schedule instead of a
12791 plain des_key_schedule (which was actually always a pointer
12794 des_key_schedule ks;
12796 des_set_key_checked(..., &ks);
12797 des_ncbc_encrypt(..., &ks, ...);
12799 (Note that a later change renames 'des_...' into 'DES_...'.)
12803 * Initial reduction of linker bloat: the use of some functions, such as
12804 PEM causes large amounts of unused functions to be linked in due to
12805 poor organisation. For example pem_all.c contains every PEM function
12806 which has a knock on effect of linking in large amounts of (unused)
12807 ASN1 code. Grouping together similar functions and splitting unrelated
12808 functions prevents this.
12812 * Cleanup of EVP macros.
12816 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12817 correct `_ecb suffix`.
12821 * Add initial OCSP responder support to ocsp application. The
12822 revocation information is handled using the text based index
12823 use by the ca application. The responder can either handle
12824 requests generated internally, supplied in files (for example
12825 via a CGI script) or using an internal minimal server.
12829 * Add configuration choices to get zlib compression for TLS.
12833 * Changes to Kerberos SSL for RFC 2712 compliance:
12834 1. Implemented real KerberosWrapper, instead of just using
12835 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12836 2. Implemented optional authenticator field of KerberosWrapper.
12838 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12839 and authenticator structs; see crypto/krb5/.
12841 Generalized Kerberos calls to support multiple Kerberos libraries.
12842 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12843 via Richard Levitte*
12845 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12846 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12847 values for each of the key sizes rather than having just
12848 parameters (and 'speed' generating keys each time).
12852 * Speed up EVP routines.
12855 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12856 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12857 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12858 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12860 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12861 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12862 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12865 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12867 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12871 * Added the OS2-EMX target.
12873 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12875 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12876 New functions to support `NCONF` routines in extension code.
12877 New function `CONF_set_nconf()`
12878 to allow functions which take an `NCONF` to also handle the old `LHASH`
12879 structure: this means that the old `CONF` compatible routines can be
12880 retained (in particular w.rt. extensions) without having to duplicate the
12881 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12885 * Enhance the general user interface with mechanisms for inner control
12886 and with possibilities to have yes/no kind of prompts.
12890 * Change all calls to low-level digest routines in the library and
12891 applications to use EVP. Add missing calls to HMAC_cleanup() and
12892 don't assume HMAC_CTX can be copied using memcpy().
12894 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12896 * Add the possibility to control engines through control names but with
12897 arbitrary arguments instead of just a string.
12898 Change the key loaders to take a UI_METHOD instead of a callback
12899 function pointer. NOTE: this breaks binary compatibility with earlier
12900 versions of OpenSSL [engine].
12901 Adapt the nCipher code for these new conditions and add a card insertion
12906 * Enhance the general user interface with mechanisms to better support
12907 dialog box interfaces, application-defined prompts, the possibility
12908 to use defaults (for example default passwords from somewhere else)
12909 and interrupts/cancellations.
12913 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12914 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12918 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12919 tidy up some unnecessarily weird code in 'sk_new()').
12921 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12923 * Change the key loading routines for ENGINEs to use the same kind
12924 callback (pem_password_cb) as all other routines that need this
12929 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12930 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12931 than this minimum value is recommended.
12935 * New random seeder for OpenVMS, using the system process statistics
12936 that are easily reachable.
12940 * Windows apparently can't transparently handle global
12941 variables defined in DLLs. Initialisations such as:
12943 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12945 won't compile. This is used by the any applications that need to
12946 declare their own ASN1 modules. This was fixed by adding the option
12947 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12948 needed for static libraries under Win32.
12952 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12953 setting of purpose and trust fields. New X509_STORE trust and
12954 purpose functions and tidy up setting in other SSL functions.
12958 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12959 structure. These are inherited by X509_STORE_CTX when it is
12960 initialised. This allows various defaults to be set in the
12961 X509_STORE structure (such as flags for CRL checking and custom
12962 purpose or trust settings) for functions which only use X509_STORE_CTX
12963 internally such as S/MIME.
12965 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12966 trust settings if they are not set in X509_STORE. This allows X509_STORE
12967 purposes and trust (in S/MIME for example) to override any set by default.
12969 Add command line options for CRL checking to smime, s_client and s_server
12974 * Initial CRL based revocation checking. If the CRL checking flag(s)
12975 are set then the CRL is looked up in the X509_STORE structure and
12976 its validity and signature checked, then if the certificate is found
12977 in the CRL the verify fails with a revoked error.
12979 Various new CRL related callbacks added to X509_STORE_CTX structure.
12981 Command line options added to 'verify' application to support this.
12983 This needs some additional work, such as being able to handle multiple
12984 CRLs with different times, extension based lookup (rather than just
12985 by subject name) and ultimately more complete V2 CRL extension
12990 * Add a general user interface API (crypto/ui/). This is designed
12991 to replace things like des_read_password and friends (backward
12992 compatibility functions using this new API are provided).
12993 The purpose is to remove prompting functions from the DES code
12994 section as well as provide for prompting through dialog boxes in
12995 a window system and the like.
12999 * Add "ex_data" support to ENGINE so implementations can add state at a
13000 per-structure level rather than having to store it globally.
13004 * Make it possible for ENGINE structures to be copied when retrieved by
13005 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13006 This causes the "original" ENGINE structure to act like a template,
13007 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13008 operational state can be localised to each ENGINE structure, despite the
13009 fact they all share the same "methods". New ENGINE structures returned in
13010 this case have no functional references and the return value is the single
13011 structural reference. This matches the single structural reference returned
13012 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13017 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13018 needs to match any other type at all we need to manually clear the
13023 * Changes to the "openssl engine" utility to include;
13024 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13025 about an ENGINE's available control commands.
13026 - executing control commands from command line arguments using the
13027 '-pre' and '-post' switches. '-post' is only used if '-t' is
13028 specified and the ENGINE is successfully initialised. The syntax for
13029 the individual commands are colon-separated, for example;
13030 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13034 * New dynamic control command support for ENGINEs. ENGINEs can now
13035 declare their own commands (numbers), names (strings), descriptions,
13036 and input types for run-time discovery by calling applications. A
13037 subset of these commands are implicitly classed as "executable"
13038 depending on their input type, and only these can be invoked through
13039 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13040 can be based on user input, config files, etc). The distinction is
13041 that "executable" commands cannot return anything other than a boolean
13042 result and can only support numeric or string input, whereas some
13043 discoverable commands may only be for direct use through
13044 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13045 pointers, or other custom uses. The "executable" commands are to
13046 support parameterisations of ENGINE behaviour that can be
13047 unambiguously defined by ENGINEs and used consistently across any
13048 OpenSSL-based application. Commands have been added to all the
13049 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13050 control over shared-library paths without source code alterations.
13054 * Changed all ENGINE implementations to dynamically allocate their
13055 ENGINEs rather than declaring them statically. Apart from this being
13056 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13057 this also allows the implementations to compile without using the
13058 internal engine_int.h header.
13062 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13063 'const' value. Any code that should be able to modify a RAND_METHOD
13064 should already have non-const pointers to it (ie. they should only
13065 modify their own ones).
13069 * Made a variety of little tweaks to the ENGINE code.
13070 - "atalla" and "ubsec" string definitions were moved from header files
13071 to C code. "nuron" string definitions were placed in variables
13072 rather than hard-coded - allowing parameterisation of these values
13073 later on via ctrl() commands.
13074 - Removed unused "#if 0"'d code.
13075 - Fixed engine list iteration code so it uses ENGINE_free() to release
13076 structural references.
13077 - Constified the RAND_METHOD element of ENGINE structures.
13078 - Constified various get/set functions as appropriate and added
13079 missing functions (including a catch-all ENGINE_cpy that duplicates
13080 all ENGINE values onto a new ENGINE except reference counts/state).
13081 - Removed NULL parameter checks in get/set functions. Setting a method
13082 or function to NULL is a way of cancelling out a previously set
13083 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13084 and doesn't justify the extra error symbols and code.
13085 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13086 flags from engine_int.h to engine.h.
13087 - Changed prototypes for ENGINE handler functions (init(), finish(),
13088 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13092 * Implement binary inversion algorithm for BN_mod_inverse in addition
13093 to the algorithm using long division. The binary algorithm can be
13094 used only if the modulus is odd. On 32-bit systems, it is faster
13095 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13096 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13097 up to 450 bits. In 64-bit environments, the binary algorithm
13098 appears to be advantageous for much longer moduli; here we use it
13099 for moduli up to 2048 bits.
13103 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13104 could not support the combine flag in choice fields.
13108 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13109 extensions from a certificate request to the certificate.
13113 * Allow multiple 'certopt' and 'nameopt' options to be separated
13114 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13115 file: this allows the display of the certificate about to be
13116 signed to be customised, to allow certain fields to be included
13117 or excluded and extension details. The old system didn't display
13118 multicharacter strings properly, omitted fields not in the policy
13119 and couldn't display additional details such as extensions.
13123 * Function EC_POINTs_mul for multiple scalar multiplication
13124 of an arbitrary number of elliptic curve points
13125 \sum scalars[i]*points[i],
13126 optionally including the generator defined for the EC_GROUP:
13127 scalar*generator + \sum scalars[i]*points[i].
13129 EC_POINT_mul is a simple wrapper function for the typical case
13130 that the point list has just one item (besides the optional
13135 * First EC_METHODs for curves over GF(p):
13137 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13138 operations and provides various method functions that can also
13139 operate with faster implementations of modular arithmetic.
13141 EC_GFp_mont_method() reuses most functions that are part of
13142 EC_GFp_simple_method, but uses Montgomery arithmetic.
13144 *Bodo Moeller; point addition and point doubling
13145 implementation directly derived from source code provided by
13146 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13148 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13149 crypto/ec/ec_lib.c):
13151 Curves are EC_GROUP objects (with an optional group generator)
13152 based on EC_METHODs that are built into the library.
13154 Points are EC_POINT objects based on EC_GROUP objects.
13156 Most of the framework would be able to handle curves over arbitrary
13157 finite fields, but as there are no obvious types for fields other
13158 than GF(p), some functions are limited to that for now.
13162 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13163 that the file contains a complete HTTP response.
13167 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13168 change the def and num file printf format specifier from "%-40sXXX"
13169 to "%-39s XXX". The latter will always guarantee a space after the
13170 field while the former will cause them to run together if the field
13171 is 40 of more characters long.
13175 * Constify the cipher and digest 'method' functions and structures
13176 and modify related functions to take constant EVP_MD and EVP_CIPHER
13181 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13182 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13186 * Modify `EVP_Digest*()` routines so they now return values. Although the
13187 internal software routines can never fail additional hardware versions
13192 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13194 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13195 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13198 ERR_R_NESTED_ASN1_ERROR
13200 ERR_R_MISSING_ASN1_EOS
13201 were 4 .. 9, conflicting with
13202 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13204 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13205 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13207 Add new error code 'ERR_R_INTERNAL_ERROR'.
13211 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13216 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13217 sets the subject name for a new request or supersedes the
13218 subject name in a given request. Formats that can be parsed are
13219 'CN=Some Name, OU=myOU, C=IT'
13221 'CN=Some Name/OU=myOU/C=IT'.
13223 Add options '-batch' and '-verbose' to 'openssl req'.
13225 *Massimiliano Pala <madwolf@hackmasters.net>*
13227 * Introduce the possibility to access global variables through
13228 functions on platform were that's the best way to handle exporting
13229 global variables in shared libraries. To enable this functionality,
13230 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13231 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13232 is normally done by Configure or something similar).
13234 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13235 in the source file (foo.c) like this:
13237 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13238 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13240 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13241 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13243 OPENSSL_DECLARE_GLOBAL(int,foo);
13244 #define foo OPENSSL_GLOBAL_REF(foo)
13245 OPENSSL_DECLARE_GLOBAL(double,bar);
13246 #define bar OPENSSL_GLOBAL_REF(bar)
13248 The #defines are very important, and therefore so is including the
13249 header file everywhere where the defined globals are used.
13251 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13252 of ASN.1 items, but that structure is a bit different.
13254 The largest change is in util/mkdef.pl which has been enhanced with
13255 better and easier to understand logic to choose which symbols should
13256 go into the Windows .def files as well as a number of fixes and code
13257 cleanup (among others, algorithm keywords are now sorted
13258 lexicographically to avoid constant rewrites).
13262 * In BN_div() keep a copy of the sign of 'num' before writing the
13263 result to 'rm' because if rm==num the value will be overwritten
13264 and produce the wrong result if 'num' is negative: this caused
13265 problems with BN_mod() and BN_nnmod().
13269 * Function OCSP_request_verify(). This checks the signature on an
13270 OCSP request and verifies the signer certificate. The signer
13271 certificate is just checked for a generic purpose and OCSP request
13276 * Add OCSP_check_validity() function to check the validity of OCSP
13277 responses. OCSP responses are prepared in real time and may only
13278 be a few seconds old. Simply checking that the current time lies
13279 between thisUpdate and nextUpdate max reject otherwise valid responses
13280 caused by either OCSP responder or client clock inaccuracy. Instead
13281 we allow thisUpdate and nextUpdate to fall within a certain period of
13282 the current time. The age of the response can also optionally be
13283 checked. Two new options -validity_period and -status_age added to
13288 * If signature or public key algorithm is unrecognized print out its
13289 OID rather that just UNKNOWN.
13293 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13294 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13295 ID to be generated from the issuer certificate alone which can then be
13296 passed to OCSP_id_issuer_cmp().
13300 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13301 ASN1 modules to export functions returning ASN1_ITEM pointers
13302 instead of the ASN1_ITEM structures themselves. This adds several
13303 new macros which allow the underlying ASN1 function/structure to
13304 be accessed transparently. As a result code should not use ASN1_ITEM
13305 references directly (such as &X509_it) but instead use the relevant
13306 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13307 use of the new ASN1 code on platforms where exporting structures
13308 is problematical (for example in shared libraries) but exporting
13309 functions returning pointers to structures is not.
13313 * Add support for overriding the generation of SSL/TLS session IDs.
13314 These callbacks can be registered either in an SSL_CTX or per SSL.
13315 The purpose of this is to allow applications to control, if they wish,
13316 the arbitrary values chosen for use as session IDs, particularly as it
13317 can be useful for session caching in multiple-server environments. A
13318 command-line switch for testing this (and any client code that wishes
13319 to use such a feature) has been added to "s_server".
13321 *Geoff Thorpe, Lutz Jaenicke*
13323 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13324 of the form `#if defined(...) || defined(...) || ...` and
13325 `#if !defined(...) && !defined(...) && ...`. This also avoids
13326 the growing number of special cases it was previously handling.
13330 * Make all configuration macros available for application by making
13331 sure they are available in opensslconf.h, by giving them names starting
13332 with `OPENSSL_` to avoid conflicts with other packages and by making
13333 sure e_os2.h will cover all platform-specific cases together with
13335 Additionally, it is now possible to define configuration/platform-
13336 specific names (called "system identities"). In the C code, these
13337 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13338 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13339 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13344 * New option -set_serial to 'req' and 'x509' this allows the serial
13345 number to use to be specified on the command line. Previously self
13346 signed certificates were hard coded with serial number 0 and the
13347 CA options of 'x509' had to use a serial number in a file which was
13352 * New options to 'ca' utility to support V2 CRL entry extensions.
13353 Currently CRL reason, invalidity date and hold instruction are
13354 supported. Add new CRL extensions to V3 code and some new objects.
13358 * New function EVP_CIPHER_CTX_set_padding() this is used to
13359 disable standard block padding (aka PKCS#5 padding) in the EVP
13360 API, which was previously mandatory. This means that the data is
13361 not padded in any way and so the total length much be a multiple
13362 of the block size, otherwise an error occurs.
13366 * Initial (incomplete) OCSP SSL support.
13370 * New function OCSP_parse_url(). This splits up a URL into its host,
13371 port and path components: primarily to parse OCSP URLs. New -url
13372 option to ocsp utility.
13376 * New nonce behavior. The return value of OCSP_check_nonce() now
13377 reflects the various checks performed. Applications can decide
13378 whether to tolerate certain situations such as an absent nonce
13379 in a response when one was present in a request: the ocsp application
13380 just prints out a warning. New function OCSP_add1_basic_nonce()
13381 this is to allow responders to include a nonce in a response even if
13382 the request is nonce-less.
13386 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13387 skipped when using openssl x509 multiple times on a single input file,
13388 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13392 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13393 set string type: to handle setting ASN1_TIME structures. Fix ca
13394 utility to correctly initialize revocation date of CRLs.
13398 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13399 the clients preferred ciphersuites and rather use its own preferences.
13400 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13401 Internet Explorer by ensuring unchanged hash method during stepup.
13402 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13406 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13407 to aes and add a new 'exist' option to print out symbols that don't
13412 * Additional options to ocsp utility to allow flags to be set and
13413 additional certificates supplied.
13417 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13418 OCSP client a number of certificate to only verify the response
13423 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13424 handle the new API. Currently only ECB, CBC modes supported. Add new
13427 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13428 Encryption Standard (AES) Ciphersuites for Transport Layer
13429 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13430 not enabled by default and were not part of the "ALL" ciphersuite
13431 alias because they were not yet official; they could be
13432 explicitly requested by specifying the "AESdraft" ciphersuite
13433 group alias. In the final release of OpenSSL 0.9.7, the group
13434 alias is called "AES" and is part of "ALL".)
13436 *Ben Laurie, Steve Henson, Bodo Moeller*
13438 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13439 request to response.
13443 * Functions for OCSP responders. OCSP_request_onereq_count(),
13444 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13445 extract information from a certificate request. OCSP_response_create()
13446 creates a response and optionally adds a basic response structure.
13447 OCSP_basic_add1_status() adds a complete single response to a basic
13448 response and returns the OCSP_SINGLERESP structure just added (to allow
13449 extensions to be included for example). OCSP_basic_add1_cert() adds a
13450 certificate to a basic response and OCSP_basic_sign() signs a basic
13451 response with various flags. New helper functions ASN1_TIME_check()
13452 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13453 (converts ASN1_TIME to GeneralizedTime).
13457 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13458 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13459 structure from a certificate. X509_pubkey_digest() digests the public_key
13460 contents: this is used in various key identifiers.
13464 * Make sk_sort() tolerate a NULL argument.
13466 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13468 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13469 passed by the function are trusted implicitly. If any of them signed the
13470 response then it is assumed to be valid and is not verified.
13474 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13475 to data. This was previously part of the PKCS7 ASN1 code. This
13476 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13477 *Steve Henson, reported by Kenneth R. Robinette
13478 <support@securenetterm.com>*
13480 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13481 routines: without these tracing memory leaks is very painful.
13482 Fix leaks in PKCS12 and PKCS7 routines.
13486 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13487 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13488 effectively meant GeneralizedTime would never be used. Now it
13489 is initialised to -1 but X509_time_adj() now has to check the value
13490 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13491 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13492 *Steve Henson, reported by Kenneth R. Robinette
13493 <support@securenetterm.com>*
13495 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13496 result in a zero length in the ASN1_INTEGER structure which was
13497 not consistent with the structure when d2i_ASN1_INTEGER() was used
13498 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13499 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13500 where it did not print out a minus for negative ASN1_INTEGER.
13504 * Add summary printout to ocsp utility. The various functions which
13505 convert status values to strings have been renamed to:
13506 OCSP_response_status_str(), OCSP_cert_status_str() and
13507 OCSP_crl_reason_str() and are no longer static. New options
13508 to verify nonce values and to disable verification. OCSP response
13509 printout format cleaned up.
13513 * Add additional OCSP certificate checks. These are those specified
13514 in RFC2560. This consists of two separate checks: the CA of the
13515 certificate being checked must either be the OCSP signer certificate
13516 or the issuer of the OCSP signer certificate. In the latter case the
13517 OCSP signer certificate must contain the OCSP signing extended key
13518 usage. This check is performed by attempting to match the OCSP
13519 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13520 in the OCSP_CERTID structures of the response.
13524 * Initial OCSP certificate verification added to OCSP_basic_verify()
13525 and related routines. This uses the standard OpenSSL certificate
13526 verify routines to perform initial checks (just CA validity) and
13527 to obtain the certificate chain. Then additional checks will be
13528 performed on the chain. Currently the root CA is checked to see
13529 if it is explicitly trusted for OCSP signing. This is used to set
13530 a root CA as a global signing root: that is any certificate that
13531 chains to that CA is an acceptable OCSP signing certificate.
13535 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13536 extensions from a separate configuration file.
13537 As when reading extensions from the main configuration file,
13538 the '-extensions ...' option may be used for specifying the
13541 *Massimiliano Pala <madwolf@comune.modena.it>*
13543 * New OCSP utility. Allows OCSP requests to be generated or
13544 read. The request can be sent to a responder and the output
13545 parsed, outputted or printed in text form. Not complete yet:
13546 still needs to check the OCSP response validity.
13550 * New subcommands for 'openssl ca':
13551 `openssl ca -status <serial>` prints the status of the cert with
13552 the given serial number (according to the index file).
13553 `openssl ca -updatedb` updates the expiry status of certificates
13556 *Massimiliano Pala <madwolf@comune.modena.it>*
13558 * New '-newreq-nodes' command option to CA.pl. This is like
13559 '-newreq', but calls 'openssl req' with the '-nodes' option
13560 so that the resulting key is not encrypted.
13562 *Damien Miller <djm@mindrot.org>*
13564 * New configuration for the GNU Hurd.
13566 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13568 * Initial code to implement OCSP basic response verify. This
13569 is currently incomplete. Currently just finds the signer's
13570 certificate and verifies the signature on the response.
13574 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13575 value of OPENSSLDIR. This is available via the new '-d' option
13576 to 'openssl version', and is also included in 'openssl version -a'.
13580 * Allowing defining memory allocation callbacks that will be given
13581 file name and line number information in additional arguments
13582 (a `const char*` and an int). The basic functionality remains, as
13583 well as the original possibility to just replace malloc(),
13584 realloc() and free() by functions that do not know about these
13585 additional arguments. To register and find out the current
13586 settings for extended allocation functions, the following
13587 functions are provided:
13589 CRYPTO_set_mem_ex_functions
13590 CRYPTO_set_locked_mem_ex_functions
13591 CRYPTO_get_mem_ex_functions
13592 CRYPTO_get_locked_mem_ex_functions
13594 These work the same way as CRYPTO_set_mem_functions and friends.
13595 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13596 extended allocation function is enabled.
13597 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13598 a conventional allocation function is enabled.
13600 *Richard Levitte, Bodo Moeller*
13602 * Finish off removing the remaining LHASH function pointer casts.
13603 There should no longer be any prototype-casting required when using
13604 the LHASH abstraction, and any casts that remain are "bugs". See
13605 the callback types and macros at the head of lhash.h for details
13606 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13610 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13611 If /dev/[u]random devices are not available or do not return enough
13612 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13614 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13615 /etc/entropy will be queried once each in this sequence, querying stops
13616 when enough entropy was collected without querying more sockets.
13620 * Change the Unix RAND_poll() variant to be able to poll several
13621 random devices, as specified by DEVRANDOM, until a sufficient amount
13622 of data has been collected. We spend at most 10 ms on each file
13623 (select timeout) and read in non-blocking mode. DEVRANDOM now
13624 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13625 (previously it was just the string "/dev/urandom"), so on typical
13626 platforms the 10 ms delay will never occur.
13627 Also separate out the Unix variant to its own file, rand_unix.c.
13628 For VMS, there's a currently-empty rand_vms.c.
13632 * Move OCSP client related routines to ocsp_cl.c. These
13633 provide utility functions which an application needing
13634 to issue a request to an OCSP responder and analyse the
13635 response will typically need: as opposed to those which an
13636 OCSP responder itself would need which will be added later.
13638 OCSP_request_sign() signs an OCSP request with an API similar
13639 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13640 response. OCSP_response_get1_basic() extracts basic response
13641 from response. OCSP_resp_find_status(): finds and extracts status
13642 information from an OCSP_CERTID structure (which will be created
13643 when the request structure is built). These are built from lower
13644 level functions which work on OCSP_SINGLERESP structures but
13645 won't normally be used unless the application wishes to examine
13646 extensions in the OCSP response for example.
13648 Replace nonce routines with a pair of functions.
13649 OCSP_request_add1_nonce() adds a nonce value and optionally
13650 generates a random value. OCSP_check_nonce() checks the
13651 validity of the nonce in an OCSP response.
13655 * Change function OCSP_request_add() to OCSP_request_add0_id().
13656 This doesn't copy the supplied OCSP_CERTID and avoids the
13657 need to free up the newly created id. Change return type
13658 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13659 This can then be used to add extensions to the request.
13660 Deleted OCSP_request_new(), since most of its functionality
13661 is now in OCSP_REQUEST_new() (and the case insensitive name
13662 clash) apart from the ability to set the request name which
13663 will be added elsewhere.
13667 * Update OCSP API. Remove obsolete extensions argument from
13668 various functions. Extensions are now handled using the new
13669 OCSP extension code. New simple OCSP HTTP function which
13670 can be used to send requests and parse the response.
13674 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13675 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13676 uses the special reorder version of SET OF to sort the attributes
13677 and reorder them to match the encoded order. This resolves a long
13678 standing problem: a verify on a PKCS7 structure just after signing
13679 it used to fail because the attribute order did not match the
13680 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13681 it uses the received order. This is necessary to tolerate some broken
13682 software that does not order SET OF. This is handled by encoding
13683 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13684 to produce the required SET OF.
13688 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13689 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13690 files to get correct declarations of the ASN.1 item variables.
13694 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13695 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13696 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13697 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13698 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13699 ASN1_ITEM and no wrapper functions.
13703 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13704 replace the old function pointer based I/O routines. Change most of
13705 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13709 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13710 lines, recognize more "algorithms" that can be deselected, and make
13711 it complain about algorithm deselection that isn't recognised.
13715 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13716 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13717 to use new functions. Add NO_ASN1_OLD which can be set to remove
13718 some old style ASN1 functions: this can be used to determine if old
13719 code will still work when these eventually go away.
13723 * New extension functions for OCSP structures, these follow the
13724 same conventions as certificates and CRLs.
13728 * New function X509V3_add1_i2d(). This automatically encodes and
13729 adds an extension. Its behaviour can be customised with various
13730 flags to append, replace or delete. Various wrappers added for
13731 certificates and CRLs.
13735 * Fix to avoid calling the underlying ASN1 print routine when
13736 an extension cannot be parsed. Correct a typo in the
13737 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13741 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13742 entries for variables.
13746 * Add functionality to `apps/openssl.c` for detecting locking
13747 problems: As the program is single-threaded, all we have
13748 to do is register a locking callback using an array for
13749 storing which locks are currently held by the program.
13753 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13754 SSL_get_ex_data_X509_STORE_idx(), which is used in
13755 ssl_verify_cert_chain() and thus can be called at any time
13756 during TLS/SSL handshakes so that thread-safety is essential.
13757 Unfortunately, the ex_data design is not at all suited
13758 for multi-threaded use, so it probably should be abolished.
13762 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13764 *Broadcom, tweaked and integrated by Geoff Thorpe*
13766 * Move common extension printing code to new function
13767 X509V3_print_extensions(). Reorganise OCSP print routines and
13768 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13772 * New function X509_signature_print() to remove duplication in some
13777 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13778 set (this was treated exactly the same as SET OF previously). This
13779 is used to reorder the STACK representing the structure to match the
13780 encoding. This will be used to get round a problem where a PKCS7
13781 structure which was signed could not be verified because the STACK
13782 order did not reflect the encoded order.
13786 * Reimplement the OCSP ASN1 module using the new code.
13790 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13791 for its ASN1 operations. The old style function pointers still exist
13792 for now but they will eventually go away.
13796 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13797 completely replaces the old ASN1 functionality with a table driven
13798 encoder and decoder which interprets an ASN1_ITEM structure describing
13799 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13800 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13801 has also been converted to the new form.
13805 * Change BN_mod_exp_recp so that negative moduli are tolerated
13806 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13807 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13808 for negative moduli.
13812 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13813 of not touching the result's sign bit.
13817 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13822 * Changed the LHASH code to use prototypes for callbacks, and created
13823 macros to declare and implement thin (optionally static) functions
13824 that provide type-safety and avoid function pointer casting for the
13825 type-specific callbacks.
13829 * Added Kerberos Cipher Suites to be used with TLS, as written in
13831 *Veers Staats <staatsvr@asc.hpc.mil>,
13832 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13834 * Reformat the FAQ so the different questions and answers can be divided
13835 in sections depending on the subject.
13839 * Have the zlib compression code load ZLIB.DLL dynamically under
13844 * New function BN_mod_sqrt for computing square roots modulo a prime
13845 (using the probabilistic Tonelli-Shanks algorithm unless
13846 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13847 be handled deterministically).
13849 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13851 * Make BN_mod_inverse faster by explicitly handling small quotients
13852 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13853 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13857 * New function BN_kronecker.
13861 * Fix BN_gcd so that it works on negative inputs; the result is
13862 positive unless both parameters are zero.
13863 Previously something reasonably close to an infinite loop was
13864 possible because numbers could be growing instead of shrinking
13865 in the implementation of Euclid's algorithm.
13869 * Fix BN_is_word() and BN_is_one() macros to take into account the
13870 sign of the number in question.
13872 Fix BN_is_word(a,w) to work correctly for w == 0.
13874 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13875 because its test if the absolute value of 'a' equals 'w'.
13876 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13877 it exists mostly for use in the implementations of BN_is_zero(),
13878 BN_is_one(), and BN_is_word().
13882 * New function BN_swap.
13886 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13887 the exponentiation functions are more likely to produce reasonable
13888 results on negative inputs.
13892 * Change BN_mod_mul so that the result is always non-negative.
13893 Previously, it could be negative if one of the factors was negative;
13894 I don't think anyone really wanted that behaviour.
13898 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13899 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13900 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13901 and add new functions:
13910 BN_mod_lshift1_quick
13912 BN_mod_lshift_quick
13914 These functions always generate non-negative results.
13916 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13917 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13919 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13920 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13921 be reduced modulo `m`.
13923 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13926 The following entry accidentally appeared in the CHANGES file
13927 distributed with OpenSSL 0.9.7. The modifications described in
13928 it do *not* apply to OpenSSL 0.9.7.
13930 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13931 was actually never needed) and in BN_mul(). The removal in BN_mul()
13932 required a small change in bn_mul_part_recursive() and the addition
13933 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13934 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13935 bn_sub_words() and bn_add_words() except they take arrays with
13941 * In 'openssl passwd', verify passwords read from the terminal
13942 unless the '-salt' option is used (which usually means that
13943 verification would just waste user's time since the resulting
13944 hash is going to be compared with some given password hash)
13945 or the new '-noverify' option is used.
13947 This is an incompatible change, but it does not affect
13948 non-interactive use of 'openssl passwd' (passwords on the command
13949 line, '-stdin' option, '-in ...' option) and thus should not
13950 cause any problems.
13954 * Remove all references to RSAref, since there's no more need for it.
13958 * Make DSO load along a path given through an environment variable
13959 (SHLIB_PATH) with shl_load().
13963 * Constify the ENGINE code as a result of BIGNUM constification.
13964 Also constify the RSA code and most things related to it. In a
13965 few places, most notable in the depth of the ASN.1 code, ugly
13966 casts back to non-const were required (to be solved at a later
13971 * Make it so the openssl application has all engines loaded by default.
13975 * Constify the BIGNUM routines a little more.
13979 * Add the following functions:
13981 ENGINE_load_cswift()
13983 ENGINE_load_atalla()
13984 ENGINE_load_nuron()
13985 ENGINE_load_builtin_engines()
13987 That way, an application can itself choose if external engines that
13988 are built-in in OpenSSL shall ever be used or not. The benefit is
13989 that applications won't have to be linked with libdl or other dso
13990 libraries unless it's really needed.
13992 Changed 'openssl engine' to load all engines on demand.
13993 Changed the engine header files to avoid the duplication of some
13994 declarations (they differed!).
13998 * 'openssl engine' can now list capabilities.
14002 * Better error reporting in 'openssl engine'.
14006 * Never call load_dh_param(NULL) in s_server.
14010 * Add engine application. It can currently list engines by name and
14011 identity, and test if they are actually available.
14015 * Improve RPM specification file by forcing symbolic linking and making
14016 sure the installed documentation is also owned by root.root.
14018 *Damien Miller <djm@mindrot.org>*
14020 * Give the OpenSSL applications more possibilities to make use of
14021 keys (public as well as private) handled by engines.
14025 * Add OCSP code that comes from CertCo.
14029 * Add VMS support for the Rijndael code.
14033 * Added untested support for Nuron crypto accelerator.
14037 * Add support for external cryptographic devices. This code was
14038 previously distributed separately as the "engine" branch.
14040 *Geoff Thorpe, Richard Levitte*
14042 * Rework the filename-translation in the DSO code. It is now possible to
14043 have far greater control over how a "name" is turned into a filename
14044 depending on the operating environment and any oddities about the
14045 different shared library filenames on each system.
14049 * Support threads on FreeBSD-elf in Configure.
14053 * Fix for SHA1 assembly problem with MASM: it produces
14054 warnings about corrupt line number information when assembling
14055 with debugging information. This is caused by the overlapping
14058 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14061 NCONF_get_number() has no error checking at all. As a replacement,
14062 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14063 promoted strongly. The old NCONF_get_number is kept around for
14064 binary backward compatibility.
14065 Make it possible for methods to load from something other than a BIO,
14066 by providing a function pointer that is given a name instead of a BIO.
14067 For example, this could be used to load configuration data from an
14072 * Fix for non blocking accept BIOs. Added new I/O special reason
14073 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14074 with non blocking I/O was not possible because no retry code was
14075 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14080 * Added the beginnings of Rijndael support.
14084 * Fix for bug in DirectoryString mask setting. Add support for
14085 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14086 to allow certificate printing to more controllable, additional
14087 'certopt' option to 'x509' to allow new printing options to be
14092 * Clean old EAY MD5 hack from e_os.h.
14096 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14098 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14099 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14101 *Joe Orton, Steve Henson*
14103 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14105 * Fix additional bug revealed by the NISCC test suite:
14107 Stop bug triggering large recursion when presented with
14108 certain ASN.1 tags ([CVE-2003-0851])
14112 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14114 * Fix various bugs revealed by running the NISCC test suite:
14116 Stop out of bounds reads in the ASN1 code when presented with
14117 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14119 If verify callback ignores invalid public key errors don't try to check
14120 certificate signature with the NULL public key.
14124 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14125 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14130 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14131 extra data after the compression methods not only for TLS 1.0
14132 but also for SSL 3.0 (as required by the specification).
14134 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14136 * Change X509_certificate_type() to mark the key as exported/exportable
14137 when it's 512 *bits* long, not 512 bytes.
14141 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14143 * Countermeasure against the Klima-Pokorny-Rosa extension of
14144 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14145 a protocol version number mismatch like a decryption error
14146 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14150 * Turn on RSA blinding by default in the default implementation
14151 to avoid a timing attack. Applications that don't want it can call
14152 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14153 They would be ill-advised to do so in most cases.
14155 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14157 * Change RSA blinding code so that it works when the PRNG is not
14158 seeded (in this case, the secret RSA exponent is abused as
14159 an unpredictable seed -- if it is not unpredictable, there
14160 is no point in blinding anyway). Make RSA blinding thread-safe
14161 by remembering the creator's thread ID in rsa->blinding and
14162 having all other threads use local one-time blinding factors
14163 (this requires more computation than sharing rsa->blinding, but
14164 avoids excessive locking; and if an RSA object is not shared
14165 between threads, blinding will still be very fast).
14169 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14171 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14172 via timing by performing a MAC computation even if incorrect
14173 block cipher padding has been found. This is a countermeasure
14174 against active attacks where the attacker has to distinguish
14175 between bad padding and a MAC verification error. ([CVE-2003-0078])
14177 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14178 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14179 Martin Vuagnoux (EPFL, Ilion)*
14181 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14183 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14184 memory from its contents. This is done with a counter that will
14185 place alternating values in each byte. This can be used to solve
14186 two issues: 1) the removal of calls to memset() by highly optimizing
14187 compilers, and 2) cleansing with other values than 0, since those can
14188 be read through on certain media, for example a swap space on disk.
14192 * Bugfix: client side session caching did not work with external caching,
14193 because the session->cipher setting was not restored when reloading
14194 from the external cache. This problem was masked, when
14195 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14196 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14200 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14201 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14203 *Zeev Lieber <zeev-l@yahoo.com>*
14205 * Undo an undocumented change introduced in 0.9.6e which caused
14206 repeated calls to OpenSSL_add_all_ciphers() and
14207 OpenSSL_add_all_digests() to be ignored, even after calling
14212 * Change the default configuration reader to deal with last line not
14213 being properly terminated.
14217 * Change X509_NAME_cmp() so it applies the special rules on handling
14218 DN values that are of type PrintableString, as well as RDNs of type
14219 emailAddress where the value has the type ia5String.
14221 *stefank@valicert.com via Richard Levitte*
14223 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14224 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14225 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14226 the bitwise-OR of the two for use by the majority of applications
14227 wanting this behaviour, and update the docs. The documented
14228 behaviour and actual behaviour were inconsistent and had been
14229 changing anyway, so this is more a bug-fix than a behavioural
14232 *Geoff Thorpe, diagnosed by Nadav Har'El*
14234 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14235 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14239 * Fix initialization code race conditions in
14240 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14241 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14242 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14243 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14244 ssl2_get_cipher_by_char(),
14245 ssl3_get_cipher_by_char().
14247 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14249 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14250 the cached sessions are flushed, as the remove_cb() might use ex_data
14251 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14252 (see [openssl.org #212]).
14254 *Geoff Thorpe, Lutz Jaenicke*
14256 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14257 length, instead of the encoding length to d2i_ASN1_OBJECT.
14261 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14263 * [In 0.9.6g-engine release:]
14264 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14266 *Lynn Gazis <lgazis@rainbow.com>*
14268 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14270 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14271 and get fix the header length calculation.
14272 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14273 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14275 * Use proper error handling instead of 'assertions' in buffer
14276 overflow checks added in 0.9.6e. This prevents DoS (the
14277 assertions could call abort()).
14279 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14281 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14283 * Add various sanity checks to asn1_get_length() to reject
14284 the ASN1 length bytes if they exceed sizeof(long), will appear
14285 negative or the content length exceeds the length of the
14288 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14290 * Fix cipher selection routines: ciphers without encryption had no flags
14291 for the cipher strength set and where therefore not handled correctly
14292 by the selection routines (PR #130).
14296 * Fix EVP_dsa_sha macro.
14301 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14302 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14303 that was added in OpenSSL 0.9.6d.
14305 As the countermeasure turned out to be incompatible with some
14306 broken SSL implementations, the new option is part of SSL_OP_ALL.
14307 SSL_OP_ALL is usually employed when compatibility with weird SSL
14308 implementations is desired (e.g. '-bugs' option to 's_client' and
14309 's_server'), so the new option is automatically set in many
14314 * Changes in security patch:
14316 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14317 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14318 Air Force Materiel Command, USAF, under agreement number
14321 * Add various sanity checks to asn1_get_length() to reject
14322 the ASN1 length bytes if they exceed sizeof(long), will appear
14323 negative or the content length exceeds the length of the
14324 supplied buffer. ([CVE-2002-0659])
14326 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14328 * Assertions for various potential buffer overflows, not known to
14329 happen in practice.
14331 *Ben Laurie (CHATS)*
14333 * Various temporary buffers to hold ASCII versions of integers were
14334 too small for 64 bit platforms. ([CVE-2002-0655])
14335 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14337 * Remote buffer overflow in SSL3 protocol - an attacker could
14338 supply an oversized session ID to a client. ([CVE-2002-0656])
14340 *Ben Laurie (CHATS)*
14342 * Remote buffer overflow in SSL2 protocol - an attacker could
14343 supply an oversized client master key. ([CVE-2002-0656])
14345 *Ben Laurie (CHATS)*
14347 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14349 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14350 encoded as NULL) with id-dsa-with-sha1.
14352 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14354 * Check various `X509_...()` return values in `apps/req.c`.
14356 *Nils Larsch <nla@trustcenter.de>*
14358 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14359 an end-of-file condition would erroneously be flagged, when the CRLF
14360 was just at the end of a processed block. The bug was discovered when
14361 processing data through a buffering memory BIO handing the data to a
14362 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14363 <ptsekov@syntrex.com> and Nedelcho Stanev.
14367 * Implement a countermeasure against a vulnerability recently found
14368 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14369 before application data chunks to avoid the use of known IVs
14370 with data potentially chosen by the attacker.
14374 * Fix length checks in ssl3_get_client_hello().
14378 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14379 to prevent ssl3_read_internal() from incorrectly assuming that
14380 ssl3_read_bytes() found application data while handshake
14381 processing was enabled when in fact s->s3->in_read_app_data was
14382 merely automatically cleared during the initial handshake.
14384 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14386 * Fix object definitions for Private and Enterprise: they were not
14387 recognized in their shortname (=lowercase) representation. Extend
14388 obj_dat.pl to issue an error when using undefined keywords instead
14389 of silently ignoring the problem (Svenning Sorensen
14390 <sss@sss.dnsalias.net>).
14394 * Fix DH_generate_parameters() so that it works for 'non-standard'
14395 generators, i.e. generators other than 2 and 5. (Previously, the
14396 code did not properly initialise the 'add' and 'rem' values to
14397 BN_generate_prime().)
14399 In the new general case, we do not insist that 'generator' is
14400 actually a primitive root: This requirement is rather pointless;
14401 a generator of the order-q subgroup is just as good, if not
14406 * Map new X509 verification errors to alerts. Discovered and submitted by
14407 Tom Wu <tom@arcot.com>.
14411 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14412 returning non-zero before the data has been completely received
14413 when using non-blocking I/O.
14415 *Bodo Moeller; problem pointed out by John Hughes*
14417 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14419 *Ben Laurie, Lutz Jaenicke*
14421 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14422 Yoram Zahavi <YoramZ@gilian.com>).
14426 * Add information about CygWin 1.3 and on, and preserve proper
14427 configuration for the versions before that.
14429 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14431 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14432 check whether we deal with a copy of a session and do not delete from
14433 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14434 <izhar@checkpoint.com>.
14438 * Do not store session data into the internal session cache, if it
14439 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14440 flag is set). Proposed by Aslam <aslam@funk.com>.
14444 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14449 * [In 0.9.6d-engine release:]
14450 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14452 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14454 * Add the configuration target linux-s390x.
14456 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14458 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14459 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14460 variable as an indication that a ClientHello message has been
14461 received. As the flag value will be lost between multiple
14462 invocations of ssl3_accept when using non-blocking I/O, the
14463 function may not be aware that a handshake has actually taken
14464 place, thus preventing a new session from being added to the
14467 To avoid this problem, we now set s->new_session to 2 instead of
14468 using a local variable.
14470 *Lutz Jaenicke, Bodo Moeller*
14472 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14473 if the SSL_R_LENGTH_MISMATCH error is detected.
14475 *Geoff Thorpe, Bodo Moeller*
14477 * New 'shared_ldflag' column in Configure platform table.
14481 * Fix EVP_CIPHER_mode macro.
14483 *"Dan S. Camper" <dan@bti.net>*
14485 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14486 type, we must throw them away by setting rr->length to 0.
14488 *D P Chang <dpc@qualys.com>*
14490 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14492 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14493 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14494 worked incorrectly for those cases where range = `10..._2` and
14495 `3*range` is two bits longer than range.)
14499 * Only add signing time to PKCS7 structures if it is not already
14504 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14505 OBJ_ld_ce should be OBJ_id_ce.
14506 Also some ip-pda OIDs in crypto/objects/objects.txt were
14507 incorrect (cf. RFC 3039).
14509 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14511 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14512 returns early because it has nothing to do.
14514 *Andy Schneider <andy.schneider@bjss.co.uk>*
14516 * [In 0.9.6c-engine release:]
14517 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14519 *Andy Schneider <andy.schneider@bjss.co.uk>*
14521 * [In 0.9.6c-engine release:]
14522 Add support for Cryptographic Appliance's keyserver technology.
14523 (Use engine 'keyclient')
14525 *Cryptographic Appliances and Geoff Thorpe*
14527 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14528 is called via tools/c89.sh because arguments have to be
14529 rearranged (all '-L' options must appear before the first object
14532 *Richard Shapiro <rshapiro@abinitio.com>*
14534 * [In 0.9.6c-engine release:]
14535 Add support for Broadcom crypto accelerator cards, backported
14538 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14540 * [In 0.9.6c-engine release:]
14541 Add support for SureWare crypto accelerator cards from
14542 Baltimore Technologies. (Use engine 'sureware')
14544 *Baltimore Technologies and Mark Cox*
14546 * [In 0.9.6c-engine release:]
14547 Add support for crypto accelerator cards from Accelerated
14548 Encryption Processing, www.aep.ie. (Use engine 'aep')
14550 *AEP Inc. and Mark Cox*
14552 * Add a configuration entry for gcc on UnixWare.
14554 *Gary Benson <gbenson@redhat.com>*
14556 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14557 messages are stored in a single piece (fixed-length part and
14558 variable-length part combined) and fix various bugs found on the way.
14562 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14563 instead. BIO_gethostbyname() does not know what timeouts are
14564 appropriate, so entries would stay in cache even when they have
14566 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14568 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14569 faced with a pathologically small ClientHello fragment that does
14570 not contain client_version: Instead of aborting with an error,
14571 simply choose the highest available protocol version (i.e.,
14572 TLS 1.0 unless it is disabled). In practice, ClientHello
14573 messages are never sent like this, but this change gives us
14574 strictly correct behaviour at least for TLS.
14578 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14579 never resets s->method to s->ctx->method when called from within
14580 one of the SSL handshake functions.
14582 *Bodo Moeller; problem pointed out by Niko Baric*
14584 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14585 (sent using the client's version number) if client_version is
14586 smaller than the protocol version in use. Also change
14587 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14588 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14589 the client will at least see that alert.
14593 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14598 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14599 client receives HelloRequest while in a handshake.
14601 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14603 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14604 should end in 'break', not 'goto end' which circumvents various
14605 cleanups done in state SSL_ST_OK. But session related stuff
14606 must be disabled for SSL_ST_OK in the case that we just sent a
14609 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14610 before just sending a HelloRequest.
14612 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14614 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14615 reveal whether illegal block cipher padding was found or a MAC
14616 verification error occurred. (Neither SSLerr() codes nor alerts
14617 are directly visible to potential attackers, but the information
14618 may leak via logfiles.)
14620 Similar changes are not required for the SSL 2.0 implementation
14621 because the number of padding bytes is sent in clear for SSL 2.0,
14622 and the extra bytes are just ignored. However ssl/s2_pkt.c
14623 failed to verify that the purported number of padding bytes is in
14628 * Add OpenUNIX-8 support including shared libraries
14629 (Boyd Lynn Gerber <gerberb@zenez.com>).
14633 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14634 'wristwatch attack' using huge encoding parameters (cf.
14635 James H. Manger's CRYPTO 2001 paper). Note that the
14636 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14637 encoding parameters and hence was not vulnerable.
14641 * BN_sqr() bug fix.
14643 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14645 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14646 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14647 followed by modular reduction.
14649 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14651 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14652 equivalent based on BN_pseudo_rand() instead of BN_rand().
14656 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14657 This function was broken, as the check for a new client hello message
14658 to handle SGC did not allow these large messages.
14659 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14663 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14667 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14668 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14672 * Rework the configuration and shared library support for Tru64 Unix.
14673 The configuration part makes use of modern compiler features and
14674 still retains old compiler behavior for those that run older versions
14675 of the OS. The shared library support part includes a variant that
14676 uses the RPATH feature, and is available through the special
14677 configuration target "alpha-cc-rpath", which will never be selected
14680 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14682 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14683 with the same message size as in ssl3_get_certificate_request().
14684 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14685 messages might inadvertently be reject as too long.
14687 *Petr Lampa <lampa@fee.vutbr.cz>*
14689 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14693 * Modified SSL library such that the verify_callback that has been set
14694 specifically for an SSL object with SSL_set_verify() is actually being
14695 used. Before the change, a verify_callback set with this function was
14696 ignored and the verify_callback() set in the SSL_CTX at the time of
14697 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14698 to allow the necessary settings.
14702 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14703 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14704 done automatically (in contradiction to the requirements of the C
14705 standard). This made problems when used from OpenSSH.
14709 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14710 dh->length and always used
14712 BN_rand_range(priv_key, dh->p).
14714 BN_rand_range() is not necessary for Diffie-Hellman, and this
14715 specific range makes Diffie-Hellman unnecessarily inefficient if
14716 dh->length (recommended exponent length) is much smaller than the
14717 length of dh->p. We could use BN_rand_range() if the order of
14718 the subgroup was stored in the DH structure, but we only have
14723 BN_rand(priv_key, l, ...)
14725 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14732 RSA_eay_public_encrypt
14733 RSA_eay_private_decrypt
14734 RSA_eay_private_encrypt (signing)
14735 RSA_eay_public_decrypt (signature verification)
14737 (default implementations for RSA_public_encrypt,
14738 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14739 always reject numbers >= n.
14743 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14744 to synchronize access to 'locking_thread'. This is necessary on
14745 systems where access to 'locking_thread' (an 'unsigned long'
14746 variable) is not atomic.
14750 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14751 *before* setting the 'crypto_lock_rand' flag. The previous code had
14752 a race condition if 0 is a valid thread ID.
14754 *Travis Vitek <vitek@roguewave.com>*
14756 * Add support for shared libraries under Irix.
14758 *Albert Chin-A-Young <china@thewrittenword.com>*
14760 * Add configuration option to build on Linux on both big-endian and
14761 little-endian MIPS.
14763 *Ralf Baechle <ralf@uni-koblenz.de>*
14765 * Add the possibility to create shared libraries on HP-UX.
14769 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14771 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14772 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14773 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14774 PRNG state recovery was possible based on the output of
14775 one PRNG request appropriately sized to gain knowledge on
14776 'md' followed by enough consecutive 1-byte PRNG requests
14777 to traverse all of 'state'.
14779 1. When updating 'md_local' (the current thread's copy of 'md')
14780 during PRNG output generation, hash all of the previous
14781 'md_local' value, not just the half used for PRNG output.
14783 2. Make the number of bytes from 'state' included into the hash
14784 independent from the number of PRNG bytes requested.
14786 The first measure alone would be sufficient to avoid
14787 Markku-Juhani's attack. (Actually it had never occurred
14788 to me that the half of 'md_local' used for chaining was the
14789 half from which PRNG output bytes were taken -- I had always
14790 assumed that the secret half would be used.) The second
14791 measure makes sure that additional data from 'state' is never
14792 mixed into 'md_local' in small portions; this heuristically
14793 further strengthens the PRNG.
14797 * Fix crypto/bn/asm/mips3.s.
14801 * When only the key is given to "enc", the IV is undefined. Print out
14802 an error message in this case.
14806 * Handle special case when X509_NAME is empty in X509 printing routines.
14810 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14811 positive and less than q.
14815 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14816 used: it isn't thread safe and the add_lock_callback should handle
14819 *Paul Rose <Paul.Rose@bridge.com>*
14821 * Verify that incoming data obeys the block size in
14822 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14828 *Ulf Möller, Bodo Möller*
14830 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14831 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14832 when fixing the server behaviour for backwards-compatible 'client
14833 hello' messages. (Note that the attack is impractical against
14834 SSL 3.0 and TLS 1.0 anyway because length and version checking
14835 means that the probability of guessing a valid ciphertext is
14836 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14839 Before 0.9.5, the countermeasure (hide the error by generating a
14840 random 'decryption result') did not work properly because
14841 ERR_clear_error() was missing, meaning that SSL_get_error() would
14842 detect the supposedly ignored error.
14844 Both problems are now fixed.
14848 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14849 (previously it was 1024).
14853 * Fix for compatibility mode trust settings: ignore trust settings
14854 unless some valid trust or reject settings are present.
14858 * Fix for blowfish EVP: its a variable length cipher.
14862 * Fix various bugs related to DSA S/MIME verification. Handle missing
14863 parameters in DSA public key structures and return an error in the
14864 DSA routines if parameters are absent.
14868 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14869 in the current directory if neither $RANDFILE nor $HOME was set.
14870 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14871 caused some confusion to Windows users who haven't defined $HOME.
14872 Thus RAND_file_name() is changed again: e_os.h can define a
14873 DEFAULT_HOME, which will be used if $HOME is not set.
14874 For Windows, we use "C:"; on other platforms, we still require
14875 environment variables.
14877 * Move 'if (!initialized) RAND_poll()' into regions protected by
14878 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14879 having multiple threads call RAND_poll() concurrently.
14883 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14884 combination of a flag and a thread ID variable.
14885 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14886 flag), *other* threads can enter ssleay_add_bytes without obeying
14887 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14888 that they do not hold after the first thread unsets add_do_not_lock).
14892 * Change bctest again: '-x' expressions are not available in all
14893 versions of 'test'.
14897 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14899 * Fix a couple of memory leaks in PKCS7_dataDecode()
14901 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14903 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14904 the default extension for executables, if any. Also, make the perl
14905 scripts that use symlink() to test if it really exists and use "cp"
14906 if it doesn't. All this made OpenSSL compilable and installable in
14911 * Fix for asn1_GetSequence() for indefinite length constructed data.
14912 If SEQUENCE is length is indefinite just set c->slen to the total
14913 amount of data available.
14915 *Steve Henson, reported by shige@FreeBSD.org*
14917 *This change does not apply to 0.9.7.*
14919 * Change bctest to avoid here-documents inside command substitution
14920 (workaround for FreeBSD /bin/sh bug).
14921 For compatibility with Ultrix, avoid shell functions (introduced
14922 in the bctest version that searches along $PATH).
14926 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14927 with des_encrypt() defined on some operating systems, like Solaris
14932 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14933 On the Importance of Eliminating Errors in Cryptographic
14934 Computations, J. Cryptology 14 (2001) 2, 101-119,
14935 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14939 * MIPS assembler BIGNUM division bug fix.
14943 * Disabled incorrect Alpha assembler code.
14947 * Fix PKCS#7 decode routines so they correctly update the length
14948 after reading an EOC for the EXPLICIT tag.
14952 *This change does not apply to 0.9.7.*
14954 * Fix bug in PKCS#12 key generation routines. This was triggered
14955 if a 3DES key was generated with a 0 initial byte. Include
14956 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14957 (but broken) behaviour.
14961 * Enhance bctest to search for a working bc along $PATH and print
14964 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14966 * Fix memory leaks in err.c: free err_data string if necessary;
14967 don't write to the wrong index in ERR_set_error_data.
14971 * Implement ssl23_peek (analogous to ssl23_read), which previously
14976 * Replace rdtsc with `_emit` statements for VC++ version 5.
14978 *Jeremy Cooper <jeremy@baymoo.org>*
14980 * Make it possible to reuse SSLv2 sessions.
14984 * In copy_email() check for >= 0 as a return value for
14985 X509_NAME_get_index_by_NID() since 0 is a valid index.
14987 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14989 * Avoid coredump with unsupported or invalid public keys by checking if
14990 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14991 PKCS7_verify() fails with non detached data.
14995 * Don't use getenv in library functions when run as setuid/setgid.
14996 New function OPENSSL_issetugid().
15000 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15001 due to incorrect handling of multi-threading:
15003 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15005 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15007 3. Count how many times MemCheck_off() has been called so that
15008 nested use can be treated correctly. This also avoids
15009 inband-signalling in the previous code (which relied on the
15010 assumption that thread ID 0 is impossible).
15014 * Add "-rand" option also to s_client and s_server.
15018 * Fix CPU detection on Irix 6.x.
15019 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15020 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15022 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15027 *This change does not apply to 0.9.7.*
15029 * Use the cached encoding of an X509_NAME structure rather than
15030 copying it. This is apparently the reason for the libsafe "errors"
15031 but the code is actually correct.
15035 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15036 Bleichenbacher's DSA attack.
15037 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15038 to be set and top=0 forces the highest bit to be set; top=-1 is new
15039 and leaves the highest bit random.
15041 *Ulf Moeller, Bodo Moeller*
15043 * In the `NCONF_...`-based implementations for `CONF_...` queries
15044 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15045 a temporary CONF structure with the data component set to NULL
15046 (which gives segmentation faults in lh_retrieve).
15047 Instead, use NULL for the CONF pointer in CONF_get_string and
15048 CONF_get_number (which may use environment variables) and directly
15049 return NULL from CONF_get_section.
15053 * Fix potential buffer overrun for EBCDIC.
15057 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15058 keyUsage if basicConstraints absent for a CA.
15062 * Make SMIME_write_PKCS7() write mail header values with a format that
15063 is more generally accepted (no spaces before the semicolon), since
15064 some programs can't parse those values properly otherwise. Also make
15065 sure BIO's that break lines after each write do not create invalid
15070 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15071 macros previously used would not encode an empty SEQUENCE OF
15072 and break the signature.
15076 *This change does not apply to 0.9.7.*
15078 * Zero the premaster secret after deriving the master secret in
15083 * Add some EVP_add_digest_alias registrations (as found in
15084 OpenSSL_add_all_digests()) to SSL_library_init()
15085 aka OpenSSL_add_ssl_algorithms(). This provides improved
15086 compatibility with peers using X.509 certificates
15087 with unconventional AlgorithmIdentifier OIDs.
15091 * Fix for Irix with NO_ASM.
15093 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15095 * ./config script fixes.
15097 *Ulf Moeller, Richard Levitte*
15099 * Fix 'openssl passwd -1'.
15103 * Change PKCS12_key_gen_asc() so it can cope with non null
15104 terminated strings whose length is passed in the passlen
15105 parameter, for example from PEM callbacks. This was done
15106 by adding an extra length parameter to asc2uni().
15108 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15110 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15111 call failed, free the DSA structure.
15115 * Fix to uni2asc() to cope with zero length Unicode strings.
15116 These are present in some PKCS#12 files.
15120 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15121 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15122 when writing a 32767 byte record.
15124 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15126 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15127 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15129 (RSA objects have a reference count access to which is protected
15130 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15131 so they are meant to be shared between threads.)
15132 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15133 "Reddie, Steven" <Steven.Reddie@ca.com>*
15135 * Fix a deadlock in CRYPTO_mem_leaks().
15139 * Use better test patterns in bntest.
15143 * rand_win.c fix for Borland C.
15147 * BN_rshift bugfix for n == 0.
15151 * Add a 'bctest' script that checks for some known 'bc' bugs
15152 so that 'make test' does not abort just because 'bc' is broken.
15156 * Store verify_result within SSL_SESSION also for client side to
15157 avoid potential security hole. (Re-used sessions on the client side
15158 always resulted in verify_result==X509_V_OK, not using the original
15159 result of the server certificate verification.)
15163 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15164 SSL3_RT_APPLICATION_DATA, return 0.
15165 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15170 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15171 releases, have been re-implemented by renaming the previous
15172 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15173 and ssl3_read_internal, respectively, and adding 'peek' parameters
15174 to them. The new ssl[23]_{read,peek} functions are calls to
15175 ssl[23]_read_internal with the 'peek' flag set appropriately.
15176 A 'peek' parameter has also been added to ssl3_read_bytes, which
15177 does the actual work for ssl3_read_internal.
15181 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15182 the method-specific "init()" handler. Also clean up ex_data after
15183 calling the method-specific "finish()" handler. Previously, this was
15184 happening the other way round.
15188 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15189 The previous value, 12, was not always sufficient for BN_mod_exp().
15193 * Make sure that shared libraries get the internal name engine with
15194 the full version number and not just 0. This should mark the
15195 shared libraries as not backward compatible. Of course, this should
15196 be changed again when we can guarantee backward binary compatibility.
15200 * Fix typo in get_cert_by_subject() in by_dir.c
15202 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15204 * Rework the system to generate shared libraries:
15206 - Make note of the expected extension for the shared libraries and
15207 if there is a need for symbolic links from for example libcrypto.so.0
15208 to libcrypto.so.0.9.7. There is extended info in Configure for
15211 - Make as few rebuilds of the shared libraries as possible.
15213 - Still avoid linking the OpenSSL programs with the shared libraries.
15215 - When installing, install the shared libraries separately from the
15220 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15222 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15223 and not in SSL_clear because the latter is also used by the
15224 accept/connect functions; previously, the settings made by
15225 SSL_set_read_ahead would be lost during the handshake.
15227 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15229 * Correct util/mkdef.pl to be selective about disabled algorithms.
15230 Previously, it would create entries for disabled algorithms no
15235 * Added several new manual pages for SSL_* function.
15239 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15241 * In ssl23_get_client_hello, generate an error message when faced
15242 with an initial SSL 3.0/TLS record that is too small to contain the
15243 first two bytes of the ClientHello message, i.e. client_version.
15244 (Note that this is a pathologic case that probably has never happened
15245 in real life.) The previous approach was to use the version number
15246 from the record header as a substitute; but our protocol choice
15247 should not depend on that one because it is not authenticated
15248 by the Finished messages.
15252 * More robust randomness gathering functions for Windows.
15254 *Jeffrey Altman <jaltman@columbia.edu>*
15256 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15257 not set then we don't setup the error code for issuer check errors
15258 to avoid possibly overwriting other errors which the callback does
15259 handle. If an application does set the flag then we assume it knows
15260 what it is doing and can handle the new informational codes
15265 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15266 a general "ANY" type, as such it should be able to decode anything
15267 including tagged types. However it didn't check the class so it would
15268 wrongly interpret tagged types in the same way as their universal
15269 counterpart and unknown types were just rejected. Changed so that the
15270 tagged and unknown types are handled in the same way as a SEQUENCE:
15271 that is the encoding is stored intact. There is also a new type
15272 "V_ASN1_OTHER" which is used when the class is not universal, in this
15273 case we have no idea what the actual type is so we just lump them all
15278 * On VMS, stdout may very well lead to a file that is written to
15279 in a record-oriented fashion. That means that every write() will
15280 write a separate record, which will be read separately by the
15281 programs trying to read from it. This can be very confusing.
15283 The solution is to put a BIO filter in the way that will buffer
15284 text until a linefeed is reached, and then write everything a
15285 line at a time, so every record written will be an actual line,
15286 not chunks of lines and not (usually doesn't happen, but I've
15287 seen it once) several lines in one record. BIO_f_linebuffer() is
15290 Currently, it's a VMS-only method, because that's where it has
15291 been tested well enough.
15295 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15296 it can return incorrect results.
15297 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15298 but it was in 0.9.6-beta[12].)
15302 * Disable the check for content being present when verifying detached
15303 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15304 include zero length content when signing messages.
15308 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15309 BIO_ctrl (for BIO pairs).
15313 * Add DSO method for VMS.
15317 * Bug fix: Montgomery multiplication could produce results with the
15322 * Add RPM specification openssl.spec and modify it to build three
15323 packages. The default package contains applications, application
15324 documentation and run-time libraries. The devel package contains
15325 include files, static libraries and function documentation. The
15326 doc package contains the contents of the doc directory. The original
15327 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15331 * Add a large number of documentation files for many SSL routines.
15333 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15335 * Add a configuration entry for Sony News 4.
15337 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15339 * Don't set the two most significant bits to one when generating a
15340 random number < q in the DSA library.
15344 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15345 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15346 the underlying transport is blocking) if a handshake took place.
15347 (The default behaviour is needed by applications such as s_client
15348 and s_server that use select() to determine when to use SSL_read;
15349 but for applications that know in advance when to expect data, it
15350 just makes things more complicated.)
15354 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15359 * Add a few more EBCDIC conditionals that make `req` and `x509`
15360 work better on such systems.
15362 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15364 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15365 Update PKCS12_parse() so it copies the friendlyName and the
15366 keyid to the certificates aux info.
15370 * Fix bug in PKCS7_verify() which caused an infinite loop
15371 if there was more than one signature.
15373 *Sven Uszpelkat <su@celocom.de>*
15375 * Major change in util/mkdef.pl to include extra information
15376 about each symbol, as well as presenting variables as well
15377 as functions. This change means that there's n more need
15378 to rebuild the .num files when some algorithms are excluded.
15382 * Allow the verify time to be set by an application,
15383 rather than always using the current time.
15387 * Phase 2 verify code reorganisation. The certificate
15388 verify code now looks up an issuer certificate by a
15389 number of criteria: subject name, authority key id
15390 and key usage. It also verifies self signed certificates
15391 by the same criteria. The main comparison function is
15392 X509_check_issued() which performs these checks.
15394 Lot of changes were necessary in order to support this
15395 without completely rewriting the lookup code.
15397 Authority and subject key identifier are now cached.
15399 The LHASH 'certs' is X509_STORE has now been replaced
15400 by a STACK_OF(X509_OBJECT). This is mainly because an
15401 LHASH can't store or retrieve multiple objects with
15402 the same hash value.
15404 As a result various functions (which were all internal
15405 use only) have changed to handle the new X509_STORE
15406 structure. This will break anything that messed round
15407 with X509_STORE internally.
15409 The functions X509_STORE_add_cert() now checks for an
15410 exact match, rather than just subject name.
15412 The X509_STORE API doesn't directly support the retrieval
15413 of multiple certificates matching a given criteria, however
15414 this can be worked round by performing a lookup first
15415 (which will fill the cache with candidate certificates)
15416 and then examining the cache for matches. This is probably
15417 the best we can do without throwing out X509_LOOKUP
15418 entirely (maybe later...).
15420 The X509_VERIFY_CTX structure has been enhanced considerably.
15422 All certificate lookup operations now go via a get_issuer()
15423 callback. Although this currently uses an X509_STORE it
15424 can be replaced by custom lookups. This is a simple way
15425 to bypass the X509_STORE hackery necessary to make this
15426 work and makes it possible to use more efficient techniques
15427 in future. A very simple version which uses a simple
15428 STACK for its trusted certificate store is also provided
15429 using X509_STORE_CTX_trusted_stack().
15431 The verify_cb() and verify() callbacks now have equivalents
15432 in the X509_STORE_CTX structure.
15434 X509_STORE_CTX also has a 'flags' field which can be used
15435 to customise the verify behaviour.
15439 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15440 excludes S/MIME capabilities.
15444 * When a certificate request is read in keep a copy of the
15445 original encoding of the signed data and use it when outputting
15446 again. Signatures then use the original encoding rather than
15447 a decoded, encoded version which may cause problems if the
15448 request is improperly encoded.
15452 * For consistency with other BIO_puts implementations, call
15453 buffer_write(b, ...) directly in buffer_puts instead of calling
15456 In BIO_puts, increment b->num_write as in BIO_write.
15458 *Peter.Sylvester@EdelWeb.fr*
15460 * Fix BN_mul_word for the case where the word is 0. (We have to use
15461 BN_zero, we may not return a BIGNUM with an array consisting of
15462 words set to zero.)
15466 * Avoid calling abort() from within the library when problems are
15467 detected, except if preprocessor symbols have been defined
15468 (such as REF_CHECK, BN_DEBUG etc.).
15472 * New openssl application 'rsautl'. This utility can be
15473 used for low-level RSA operations. DER public key
15474 BIO/fp routines also added.
15478 * New Configure entry and patches for compiling on QNX 4.
15480 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15482 * A demo state-machine implementation was sponsored by
15483 Nuron (<http://www.nuron.com/>) and is now available in
15484 demos/state_machine.
15488 * New options added to the 'dgst' utility for signature
15489 generation and verification.
15493 * Unrecognized PKCS#7 content types are now handled via a
15494 catch all ASN1_TYPE structure. This allows unsupported
15495 types to be stored as a "blob" and an application can
15496 encode and decode it manually.
15500 * Fix various signed/unsigned issues to make a_strex.c
15501 compile under VC++.
15503 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15505 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15506 length if passed a buffer. ASN1_INTEGER_to_BN failed
15507 if passed a NULL BN and its argument was negative.
15509 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15511 * Modification to PKCS#7 encoding routines to output definite
15512 length encoding. Since currently the whole structures are in
15513 memory there's not real point in using indefinite length
15514 constructed encoding. However if OpenSSL is compiled with
15515 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15519 * Added BIO_vprintf() and BIO_vsnprintf().
15523 * Added more prefixes to parse for in the strings written
15524 through a logging bio, to cover all the levels that are available
15525 through syslog. The prefixes are now:
15527 PANIC, EMERG, EMR => LOG_EMERG
15528 ALERT, ALR => LOG_ALERT
15529 CRIT, CRI => LOG_CRIT
15530 ERROR, ERR => LOG_ERR
15531 WARNING, WARN, WAR => LOG_WARNING
15532 NOTICE, NOTE, NOT => LOG_NOTICE
15533 INFO, INF => LOG_INFO
15534 DEBUG, DBG => LOG_DEBUG
15536 and as before, if none of those prefixes are present at the
15537 beginning of the string, LOG_ERR is chosen.
15539 On Win32, the `LOG_*` levels are mapped according to this:
15541 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15542 LOG_WARNING => EVENTLOG_WARNING_TYPE
15543 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15547 * Made it possible to reconfigure with just the configuration
15548 argument "reconf" or "reconfigure". The command line arguments
15549 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15550 and are retrieved from there when reconfiguring.
15556 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15558 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15562 * The obj_dat.pl script was messing up the sorting of object
15563 names. The reason was that it compared the quoted version
15564 of strings as a result "OCSP" > "OCSP Signing" because
15565 " > SPACE. Changed script to store unquoted versions of
15566 names and add quotes on output. It was also omitting some
15567 names from the lookup table if they were given a default
15568 value (that is if SN is missing it is given the same
15569 value as LN and vice versa), these are now added on the
15570 grounds that if an object has a name we should be able to
15571 look it up. Finally added warning output when duplicate
15572 short or long names are found.
15576 * Changes needed for Tandem NSK.
15578 *Scott Uroff <scott@xypro.com>*
15580 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15581 RSA_padding_check_SSLv23(), special padding was never detected
15582 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15583 version rollback attacks was not effective.
15585 In s23_clnt.c, don't use special rollback-attack detection padding
15586 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15587 client; similarly, in s23_srvr.c, don't do the rollback check if
15588 SSL 2.0 is the only protocol enabled in the server.
15592 * Make it possible to get hexdumps of unprintable data with 'openssl
15593 asn1parse'. By implication, the functions ASN1_parse_dump() and
15594 BIO_dump_indent() are added.
15598 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15599 these print out strings and name structures based on various
15600 flags including RFC2253 support and proper handling of
15601 multibyte characters. Added options to the 'x509' utility
15602 to allow the various flags to be set.
15606 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15607 Also change the functions X509_cmp_current_time() and
15608 X509_gmtime_adj() work with an ASN1_TIME structure,
15609 this will enable certificates using GeneralizedTime in validity
15610 dates to be checked.
15614 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15615 negative public key encodings) on by default,
15616 NO_NEG_PUBKEY_BUG can be set to disable it.
15620 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15621 content octets. An i2c_ASN1_OBJECT is unnecessary because
15622 the encoding can be trivially obtained from the structure.
15626 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15627 not read locks (`CRYPTO_r_[un]lock`).
15631 * A first attempt at creating official support for shared
15632 libraries through configuration. I've kept it so the
15633 default is static libraries only, and the OpenSSL programs
15634 are always statically linked for now, but there are
15635 preparations for dynamic linking in place.
15636 This has been tested on Linux and Tru64.
15640 * Randomness polling function for Win9x, as described in:
15641 Peter Gutmann, Software Generation of Practically Strong
15646 * Fix so PRNG is seeded in req if using an already existing
15651 * New options to smime application. -inform and -outform
15652 allow alternative formats for the S/MIME message including
15653 PEM and DER. The -content option allows the content to be
15654 specified separately. This should allow things like Netscape
15655 form signing output easier to verify.
15659 * Fix the ASN1 encoding of tags using the 'long form'.
15663 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15664 STRING types. These convert content octets to and from the
15665 underlying type. The actual tag and length octets are
15666 already assumed to have been read in and checked. These
15667 are needed because all other string types have virtually
15668 identical handling apart from the tag. By having versions
15669 of the ASN1 functions that just operate on content octets
15670 IMPLICIT tagging can be handled properly. It also allows
15671 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15672 and ASN1_INTEGER are identical apart from the tag.
15676 * Change the handling of OID objects as follows:
15678 - New object identifiers are inserted in objects.txt, following
15679 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15680 - objects.pl is used to process obj_mac.num and create a new
15682 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15685 This is currently kind of a hack, and the perl code in objects.pl
15686 isn't very elegant, but it works as I intended. The simplest way
15687 to check that it worked correctly is to look in obj_dat.h and
15688 check the array nid_objs and make sure the objects haven't moved
15689 around (this is important!). Additions are OK, as well as
15690 consistent name changes.
15694 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15698 * Addition of the command line parameter '-rand file' to 'openssl req'.
15699 The given file adds to whatever has already been seeded into the
15700 random pool through the RANDFILE configuration file option or
15701 environment variable, or the default random state file.
15705 * mkstack.pl now sorts each macro group into lexical order.
15706 Previously the output order depended on the order the files
15707 appeared in the directory, resulting in needless rewriting
15712 * Patches to make OpenSSL compile under Win32 again. Mostly
15713 work arounds for the VC++ problem that it treats func() as
15714 func(void). Also stripped out the parts of mkdef.pl that
15715 added extra typesafe functions: these no longer exist.
15719 * Reorganisation of the stack code. The macros are now all
15720 collected in safestack.h . Each macro is defined in terms of
15721 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15722 DEBUG_SAFESTACK is now handled in terms of function casts,
15723 this has the advantage of retaining type safety without the
15724 use of additional functions. If DEBUG_SAFESTACK is not defined
15725 then the non typesafe macros are used instead. Also modified the
15726 mkstack.pl script to handle the new form. Needs testing to see
15727 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15728 the default if no major problems. Similar behaviour for ASN1_SET_OF
15729 and PKCS12_STACK_OF.
15733 * When some versions of IIS use the 'NET' form of private key the
15734 key derivation algorithm is different. Normally MD5(password) is
15735 used as a 128 bit RC4 key. In the modified case
15736 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15737 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15738 as the old Netscape_RSA functions except they have an additional
15739 'sgckey' parameter which uses the modified algorithm. Also added
15740 an -sgckey command line option to the rsa utility. Thanks to
15741 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15742 algorithm to openssl-dev.
15746 * The evp_local.h macros were using 'c.##kname' which resulted in
15747 invalid expansion on some systems (SCO 5.0.5 for example).
15748 Corrected to 'c.kname'.
15750 *Phillip Porch <root@theporch.com>*
15752 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15753 a STACK of email addresses from a certificate or request, these look
15754 in the subject name and the subject alternative name extensions and
15755 omit any duplicate addresses.
15759 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15760 This makes DSA verification about 2 % faster.
15764 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15765 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15766 plus overhead for 1024 bit moduli).
15767 This makes exponentiations about 0.5 % faster for 1024 bit
15768 exponents (as measured by "openssl speed rsa2048").
15772 * Rename memory handling macros to avoid conflicts with other
15774 Malloc => OPENSSL_malloc
15775 Malloc_locked => OPENSSL_malloc_locked
15776 Realloc => OPENSSL_realloc
15777 Free => OPENSSL_free
15781 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15782 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15786 * CygWin32 support.
15788 *John Jarvie <jjarvie@newsguy.com>*
15790 * The type-safe stack code has been rejigged. It is now only compiled
15791 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15792 by default all type-specific stack functions are "#define"d back to
15793 standard stack functions. This results in more streamlined output
15794 but retains the type-safety checking possibilities of the original
15799 * The STACK code has been cleaned up, and certain type declarations
15800 that didn't make a lot of sense have been brought in line. This has
15801 also involved a cleanup of sorts in safestack.h to more correctly
15802 map type-safe stack functions onto their plain stack counterparts.
15803 This work has also resulted in a variety of "const"ifications of
15804 lots of the code, especially `_cmp` operations which should normally
15805 be prototyped with "const" parameters anyway.
15809 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15810 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15811 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15812 where all of 'md' is used each time the PRNG is used, but 'state'
15813 is used only indexed by a cyclic counter. As entropy may not be
15814 well distributed from the beginning, 'md' is important as a
15815 chaining variable. However, the output function chains only half
15816 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15817 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15818 in all of 'state' being rewritten, with the new values depending
15819 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15823 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15824 the handshake is continued after ssl_verify_cert_chain();
15825 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15826 can lead to 'unexplainable' connection aborts later.
15828 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15830 * Major EVP API cipher revision.
15831 Add hooks for extra EVP features. This allows various cipher
15832 parameters to be set in the EVP interface. Support added for variable
15833 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15834 setting of RC2 and RC5 parameters.
15836 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15839 Remove lots of duplicated code from the EVP library. For example *every*
15840 cipher init() function handles the 'iv' in the same way according to the
15841 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15842 for CFB and OFB modes they zero ctx->num.
15844 New functionality allows removal of S/MIME code RC2 hack.
15846 Most of the routines have the same form and so can be declared in terms
15849 By shifting this to the top level EVP_CipherInit() it can be removed from
15850 all individual ciphers. If the cipher wants to handle IVs or keys
15851 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15854 Change lots of functions like EVP_EncryptUpdate() to now return a
15855 value: although software versions of the algorithms cannot fail
15856 any installed hardware versions can.
15860 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15861 this option is set, tolerate broken clients that send the negotiated
15862 protocol version number instead of the requested protocol version
15867 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15868 i.e. non-zero for export ciphersuites, zero otherwise.
15869 Previous versions had this flag inverted, inconsistent with
15870 rsa_tmp_cb (..._TMP_RSA_CB).
15872 *Bodo Moeller; problem reported by Amit Chopra*
15874 * Add missing DSA library text string. Work around for some IIS
15875 key files with invalid SEQUENCE encoding.
15879 * Add a document (doc/standards.txt) that list all kinds of standards
15880 and so on that are implemented in OpenSSL.
15884 * Enhance c_rehash script. Old version would mishandle certificates
15885 with the same subject name hash and wouldn't handle CRLs at all.
15886 Added -fingerprint option to crl utility, to support new c_rehash
15891 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15895 * Fix for SSL server purpose checking. Server checking was
15896 rejecting certificates which had extended key usage present
15897 but no ssl client purpose.
15899 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15901 * Make PKCS#12 code work with no password. The PKCS#12 spec
15902 is a little unclear about how a blank password is handled.
15903 Since the password in encoded as a BMPString with terminating
15904 double NULL a zero length password would end up as just the
15905 double NULL. However no password at all is different and is
15906 handled differently in the PKCS#12 key generation code. NS
15907 treats a blank password as zero length. MSIE treats it as no
15908 password on export: but it will try both on import. We now do
15909 the same: PKCS12_parse() tries zero length and no password if
15910 the password is set to "" or NULL (NULL is now a valid password:
15911 it wasn't before) as does the pkcs12 application.
15915 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15916 perror when PEM_read_bio_X509_REQ fails, the error message must
15917 be obtained from the error queue.
15921 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15922 it in ERR_remove_state if appropriate, and change ERR_get_state
15923 accordingly to avoid race conditions (this is necessary because
15924 thread_hash is no longer constant once set).
15928 * Bugfix for linux-elf makefile.one.
15932 * RSA_get_default_method() will now cause a default
15933 RSA_METHOD to be chosen if one doesn't exist already.
15934 Previously this was only set during a call to RSA_new()
15935 or RSA_new_method(NULL) meaning it was possible for
15936 RSA_get_default_method() to return NULL.
15940 * Added native name translation to the existing DSO code
15941 that will convert (if the flag to do so is set) filenames
15942 that are sufficiently small and have no path information
15943 into a canonical native form. Eg. "blah" converted to
15944 "libblah.so" or "blah.dll" etc.
15948 * New function ERR_error_string_n(e, buf, len) which is like
15949 ERR_error_string(e, buf), but writes at most 'len' bytes
15950 including the 0 terminator. For ERR_error_string_n, 'buf'
15953 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15955 * CONF library reworked to become more general. A new CONF
15956 configuration file reader "class" is implemented as well as a
15957 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15958 old `CONF_*` functions are still there, but are reimplemented to
15959 work in terms of the new functions. Also, a set of functions
15960 to handle the internal storage of the configuration data is
15961 provided to make it easier to write new configuration file
15962 reader "classes" (I can definitely see something reading a
15963 configuration file in XML format, for example), called `_CONF_*`,
15964 or "the configuration storage API"...
15966 The new configuration file reading functions are:
15968 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15969 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15971 NCONF_default, NCONF_WIN32
15973 NCONF_dump_fp, NCONF_dump_bio
15975 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15976 NCONF_new creates a new CONF object. This works in the same way
15977 as other interfaces in OpenSSL, like the BIO interface.
15978 `NCONF_dump_*` dump the internal storage of the configuration file,
15979 which is useful for debugging. All other functions take the same
15980 arguments as the old `CONF_*` functions with the exception of the
15981 first that must be a `CONF *` instead of a `LHASH *`.
15983 To make it easier to use the new classes with the old `CONF_*` functions,
15984 the function CONF_set_default_method is provided.
15988 * Add '-tls1' option to 'openssl ciphers', which was already
15989 mentioned in the documentation but had not been implemented.
15990 (This option is not yet really useful because even the additional
15991 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15995 * Initial DSO code added into libcrypto for letting OpenSSL (and
15996 OpenSSL-based applications) load shared libraries and bind to
15997 them in a portable way.
15999 *Geoff Thorpe, with contributions from Richard Levitte*
16001 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16003 * Make sure _lrotl and _lrotr are only used with MSVC.
16005 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16006 (the default implementation of RAND_status).
16008 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16009 to '-clrext' (= clear extensions), as intended and documented.
16010 *Bodo Moeller; inconsistency pointed out by Michael Attili
16011 <attili@amaxo.com>*
16013 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16014 was larger than the MD block size.
16016 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16018 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16019 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16020 using the passed key: if the passed key was a private key the result
16021 of X509_print(), for example, would be to print out all the private key
16026 * des_quad_cksum() byte order bug fix.
16027 *Ulf Möller, using the problem description in krb4-0.9.7, where
16028 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16030 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16033 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16035 * For easily testing in shell scripts whether some command
16036 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16037 returns with exit code 0 iff no command of the given name is available.
16038 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16039 the output goes to stdout and nothing is printed to stderr.
16040 Additional arguments are always ignored.
16042 Since for each cipher there is a command of the same name,
16043 the 'no-cipher' compilation switches can be tested this way.
16045 ('openssl no-XXX' is not able to detect pseudo-commands such
16046 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16050 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16054 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16055 is set; it will be thrown away anyway because each handshake creates
16057 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16058 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16059 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16060 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16064 * New s_client option -ign_eof: EOF at stdin is ignored, and
16065 'Q' and 'R' lose their special meanings (quit/renegotiate).
16066 This is part of what -quiet does; unlike -quiet, -ign_eof
16067 does not suppress any output.
16071 * Add compatibility options to the purpose and trust code. The
16072 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16073 accepts a certificate or CA, this was the previous behaviour,
16074 with all the associated security issues.
16076 X509_TRUST_COMPAT is the old trust behaviour: only and
16077 automatically trust self signed roots in certificate store. A
16078 new trust setting X509_TRUST_DEFAULT is used to specify that
16079 a purpose has no associated trust setting and it should instead
16080 use the value in the default purpose.
16084 * Fix the PKCS#8 DSA private key code so it decodes keys again
16085 and fix a memory leak.
16089 * In util/mkerr.pl (which implements 'make errors'), preserve
16090 reason strings from the previous version of the .c file, as
16091 the default to have only downcase letters (and digits) in
16092 automatically generated reasons codes is not always appropriate.
16096 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16097 using strerror. Previously, ERR_reason_error_string() returned
16098 library names as reason strings for SYSerr; but SYSerr is a special
16099 case where small numbers are errno values, not library numbers.
16103 * Add '-dsaparam' option to 'openssl dhparam' application. This
16104 converts DSA parameters into DH parameters. (When creating parameters,
16105 DSA_generate_parameters is used.)
16109 * Include 'length' (recommended exponent length) in C code generated
16110 by 'openssl dhparam -C'.
16114 * The second argument to set_label in perlasm was already being used
16115 so couldn't be used as a "file scope" flag. Moved to third argument
16120 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16121 instead of RAND_bytes for encryption IVs and salts.
16125 * Include RAND_status() into RAND_METHOD instead of implementing
16126 it only for md_rand.c Otherwise replacing the PRNG by calling
16127 RAND_set_rand_method would be impossible.
16131 * Don't let DSA_generate_key() enter an infinite loop if the random
16132 number generation fails.
16136 * New 'rand' application for creating pseudo-random output.
16140 * Added configuration support for Linux/IA64
16142 *Rolf Haberrecker <rolf@suse.de>*
16144 * Assembler module support for Mingw32.
16148 * Shared library support for HPUX (in shlib/).
16150 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16152 * Shared library support for Solaris gcc.
16154 *Lutz Behnke <behnke@trustcenter.de>*
16156 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16158 * PKCS7_encrypt() was adding text MIME headers twice because they
16159 were added manually and by SMIME_crlf_copy().
16163 * In bntest.c don't call BN_rand with zero bits argument.
16165 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16167 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16168 case was implemented. This caused BN_div_recp() to fail occasionally.
16172 * Add an optional second argument to the set_label() in the perl
16173 assembly language builder. If this argument exists and is set
16174 to 1 it signals that the assembler should use a symbol whose
16175 scope is the entire file, not just the current function. This
16176 is needed with MASM which uses the format label:: for this scope.
16178 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16180 * Change the ASN1 types so they are typedefs by default. Before
16181 almost all types were #define'd to ASN1_STRING which was causing
16182 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16187 * Change names of new functions to the new get1/get0 naming
16188 convention: After 'get1', the caller owns a reference count
16189 and has to call `..._free`; 'get0' returns a pointer to some
16190 data structure without incrementing reference counters.
16191 (Some of the existing 'get' functions increment a reference
16192 counter, some don't.)
16193 Similarly, 'set1' and 'add1' functions increase reference
16194 counters or duplicate objects.
16198 * Allow for the possibility of temp RSA key generation failure:
16199 the code used to assume it always worked and crashed on failure.
16203 * Fix potential buffer overrun problem in BIO_printf().
16204 *Ulf Möller, using public domain code by Patrick Powell; problem
16205 pointed out by David Sacerdote <das33@cornell.edu>*
16207 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16208 RAND_egd() and RAND_status(). In the command line application,
16209 the EGD socket can be specified like a seed file using RANDFILE
16214 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16215 Some CAs (e.g. Verisign) distribute certificates in this form.
16219 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16220 list to exclude them. This means that no special compilation option
16221 is needed to use anonymous DH: it just needs to be included in the
16226 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16227 EVP_MD_type. The old functionality is available in a new macro called
16228 EVP_MD_md(). Change code that uses it and update docs.
16232 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16233 where the `void *` argument is replaced by a function pointer argument.
16234 Previously `void *` was abused to point to functions, which works on
16235 many platforms, but is not correct. As these functions are usually
16236 called by macros defined in OpenSSL header files, most source code
16237 should work without changes.
16241 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16242 sections with information on -D... compiler switches used for
16243 compiling the library so that applications can see them. To enable
16244 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16245 must be defined. E.g.,
16246 #define OPENSSL_ALGORITHM_DEFINES
16247 #include <openssl/opensslconf.h>
16248 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16250 *Richard Levitte, Ulf and Bodo Möller*
16252 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16257 * Change the 'other' type in certificate aux info to a STACK_OF
16258 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16259 the required ASN1 format: arbitrary types determined by an OID.
16263 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16264 argument to 'req'. This is not because the function is newer or
16265 better than others it just uses the work 'NEW' in the certificate
16266 request header lines. Some software needs this.
16270 * Reorganise password command line arguments: now passwords can be
16271 obtained from various sources. Delete the PEM_cb function and make
16272 it the default behaviour: i.e. if the callback is NULL and the
16273 usrdata argument is not NULL interpret it as a null terminated pass
16274 phrase. If usrdata and the callback are NULL then the pass phrase
16275 is prompted for as usual.
16279 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16280 the support is automatically enabled. The resulting binaries will
16281 autodetect the card and use it if present.
16283 *Ben Laurie and Compaq Inc.*
16285 * Work around for Netscape hang bug. This sends certificate request
16286 and server done in one record. Since this is perfectly legal in the
16287 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16288 the bugs/SSLv3 entry for more info.
16292 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16296 * Add -rand argument to smime and pkcs12 applications and read/write
16301 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16305 * Add command line password options to the remaining applications.
16309 * Bug fix for BN_div_recp() for numerators with an even number of
16314 * More tests in bntest.c, and changed test_bn output.
16318 * ./config recognizes MacOS X now.
16322 * Bug fix for BN_div() when the first words of num and divisor are
16323 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16327 * Add support for various broken PKCS#8 formats, and command line
16328 options to produce them.
16332 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16333 get temporary BIGNUMs from a BN_CTX.
16337 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16342 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16343 include a #define from the old name to the new. The original intent
16344 was that statically linked binaries could for example just call
16345 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16346 link with digests. This never worked because SSLeay_add_all_digests()
16347 and SSLeay_add_all_ciphers() were in the same source file so calling
16348 one would link with the other. They are now in separate source files.
16352 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16356 * Use a less unusual form of the Miller-Rabin primality test (it used
16357 a binary algorithm for exponentiation integrated into the Miller-Rabin
16358 loop, our standard modexp algorithms are faster).
16362 * Support for the EBCDIC character set completed.
16364 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16366 * Source code cleanups: use const where appropriate, eliminate casts,
16367 use `void *` instead of `char *` in lhash.
16371 * Bugfix: ssl3_send_server_key_exchange was not restartable
16372 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16373 this the server could overwrite ephemeral keys that the client
16378 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16379 using 50 iterations of the Rabin-Miller test.
16381 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16382 iterations of the Rabin-Miller test as required by the appendix
16383 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16384 As BN_is_prime_fasttest includes trial division, DSA parameter
16385 generation becomes much faster.
16387 This implies a change for the callback functions in DSA_is_prime
16388 and DSA_generate_parameters: The callback function is called once
16389 for each positive witness in the Rabin-Miller test, not just
16390 occasionally in the inner loop; and the parameters to the
16391 callback function now provide an iteration count for the outer
16392 loop rather than for the current invocation of the inner loop.
16393 DSA_generate_parameters additionally can call the callback
16394 function with an 'iteration count' of -1, meaning that a
16395 candidate has passed the trial division test (when q is generated
16396 from an application-provided seed, trial division is skipped).
16400 * New function BN_is_prime_fasttest that optionally does trial
16401 division before starting the Rabin-Miller test and has
16402 an additional BN_CTX * argument (whereas BN_is_prime always
16403 has to allocate at least one BN_CTX).
16404 'callback(1, -1, cb_arg)' is called when a number has passed the
16405 trial division stage.
16409 * Fix for bug in CRL encoding. The validity dates weren't being handled
16414 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16418 * New function BN_pseudo_rand().
16422 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16423 bignum version of BN_from_montgomery() with the working code from
16424 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16429 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16430 made it impossible to use the same SSL_SESSION data structure in
16431 SSL2 clients in multiple threads.
16435 * The return value of RAND_load_file() no longer counts bytes obtained
16436 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16437 to seed the PRNG (previously an explicit byte count was required).
16439 *Ulf Möller, Bodo Möller*
16441 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16442 used `char *` instead of `void *` and had casts all over the place.
16446 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16450 * Retain source code compatibility for BN_prime_checks macro:
16451 BN_is_prime(..., BN_prime_checks, ...) now uses
16452 BN_prime_checks_for_size to determine the appropriate number of
16453 Rabin-Miller iterations.
16457 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16458 DH_CHECK_P_NOT_SAFE_PRIME.
16459 (Check if this is true? OpenPGP calls them "strong".)
16463 * Merge the functionality of "dh" and "gendh" programs into a new program
16464 "dhparam". The old programs are retained for now but will handle DH keys
16465 (instead of parameters) in future.
16469 * Make the ciphers, s_server and s_client programs check the return values
16470 when a new cipher list is set.
16474 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16475 ciphers. Before when the 56bit ciphers were enabled the sorting was
16478 The syntax for the cipher sorting has been extended to support sorting by
16479 cipher-strength (using the strength_bits hard coded in the tables).
16480 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16482 Fix a bug in the cipher-command parser: when supplying a cipher command
16483 string with an "undefined" symbol (neither command nor alphanumeric
16484 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16485 an error is flagged.
16487 Due to the strength-sorting extension, the code of the
16488 ssl_create_cipher_list() function was completely rearranged. I hope that
16489 the readability was also increased :-)
16491 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16493 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16494 for the first serial number and places 2 in the serial number file. This
16495 avoids problems when the root CA is created with serial number zero and
16496 the first user certificate has the same issuer name and serial number
16501 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16502 the new code. Add documentation for this stuff.
16506 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16507 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16508 structures and behave in an analogous way to the X509v3 functions:
16509 they shouldn't be called directly but wrapper functions should be used
16512 So we also now have some wrapper functions that call the X509at functions
16513 when passed certificate requests. (TO DO: similar things can be done with
16514 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16515 things. Some of these need some d2i or i2d and print functionality
16516 because they handle more complex structures.)
16520 * Add missing #ifndefs that caused missing symbols when building libssl
16521 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16522 NO_RSA in `ssl/s2*.c`.
16524 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16526 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16527 has a return value which indicates the quality of the random data
16528 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16529 error queue. New function RAND_pseudo_bytes() generates output that is
16530 guaranteed to be unique but not unpredictable. RAND_add is like
16531 RAND_seed, but takes an extra argument for an entropy estimate
16532 (RAND_seed always assumes full entropy).
16536 * Do more iterations of Rabin-Miller probable prime test (specifically,
16537 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16538 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16539 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16540 false-positive rate of at most 2^-80 for random input.
16544 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16548 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16549 in the 0.9.5 release), this returns the chain
16550 from an X509_CTX structure with a dup of the stack and all
16551 the X509 reference counts upped: so the stack will exist
16552 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16555 Also make SSL_SESSION_print() print out the verify return
16560 * Add manpage for the pkcs12 command. Also change the default
16561 behaviour so MAC iteration counts are used unless the new
16562 -nomaciter option is used. This improves file security and
16563 only older versions of MSIE (4.0 for example) need it.
16567 * Honor the no-xxx Configure options when creating .DEF files.
16571 * Add PKCS#10 attributes to field table: challengePassword,
16572 unstructuredName and unstructuredAddress. These are taken from
16573 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16574 international characters are used.
16576 More changes to X509_ATTRIBUTE code: allow the setting of types
16577 based on strings. Remove the 'loc' parameter when adding
16578 attributes because these will be a SET OF encoding which is sorted
16583 * Initial changes to the 'req' utility to allow request generation
16584 automation. This will allow an application to just generate a template
16585 file containing all the field values and have req construct the
16588 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16589 used all over the place including certificate requests and PKCS#7
16590 structures. They are currently handled manually where necessary with
16591 some primitive wrappers for PKCS#7. The new functions behave in a
16592 manner analogous to the X509 extension functions: they allow
16593 attributes to be looked up by NID and added.
16595 Later something similar to the X509V3 code would be desirable to
16596 automatically handle the encoding, decoding and printing of the
16597 more complex types. The string types like challengePassword can
16598 be handled by the string table functions.
16600 Also modified the multi byte string table handling. Now there is
16601 a 'global mask' which masks out certain types. The table itself
16602 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16603 is useful when for example there is only one permissible type
16604 (as in countryName) and using the mask might result in no valid
16609 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16610 SSL_get_peer_finished to allow applications to obtain the latest
16611 Finished messages sent to the peer or expected from the peer,
16612 respectively. (SSL_get_peer_finished is usually the Finished message
16613 actually received from the peer, otherwise the protocol will be aborted.)
16615 As the Finished message are message digests of the complete handshake
16616 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16617 be used for external authentication procedures when the authentication
16618 provided by SSL/TLS is not desired or is not enough.
16622 * Enhanced support for Alpha Linux is added. Now ./config checks if
16623 the host supports BWX extension and if Compaq C is present on the
16624 $PATH. Just exploiting of the BWX extension results in 20-30%
16625 performance kick for some algorithms, e.g. DES and RC4 to mention
16626 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16631 * Add support for MS "fast SGC". This is arguably a violation of the
16632 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16633 weak crypto and after checking the certificate is SGC a second one
16634 with strong crypto. MS SGC stops the first handshake after receiving
16635 the server certificate message and sends a second client hello. Since
16636 a server will typically do all the time consuming operations before
16637 expecting any further messages from the client (server key exchange
16638 is the most expensive) there is little difference between the two.
16640 To get OpenSSL to support MS SGC we have to permit a second client
16641 hello message after we have sent server done. In addition we have to
16642 reset the MAC if we do get this second client hello.
16646 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16647 if a DER encoded private key is RSA or DSA traditional format. Changed
16648 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16649 format DER encoded private key. Newer code should use PKCS#8 format which
16650 has the key type encoded in the ASN1 structure. Added DER private key
16651 support to pkcs8 application.
16655 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16656 ciphersuites has been selected (as required by the SSL 3/TLS 1
16657 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16658 is set, we interpret this as a request to violate the specification
16659 (the worst that can happen is a handshake failure, and 'correct'
16660 behaviour would result in a handshake failure anyway).
16664 * In SSL_CTX_add_session, take into account that there might be multiple
16665 SSL_SESSION structures with the same session ID (e.g. when two threads
16666 concurrently obtain them from an external cache).
16667 The internal cache can handle only one SSL_SESSION with a given ID,
16668 so if there's a conflict, we now throw out the old one to achieve
16673 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16674 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16675 some routines that use cipher OIDs: some ciphers do not have OIDs
16676 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16681 * Simplify the trust setting structure and code. Now we just have
16682 two sequences of OIDs for trusted and rejected settings. These will
16683 typically have values the same as the extended key usage extension
16684 and any application specific purposes.
16686 The trust checking code now has a default behaviour: it will just
16687 check for an object with the same NID as the passed id. Functions can
16688 be provided to override either the default behaviour or the behaviour
16689 for a given id. SSL client, server and email already have functions
16690 in place for compatibility: they check the NID and also return "trusted"
16691 if the certificate is self signed.
16695 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16696 traditional format into an EVP_PKEY structure.
16700 * Add a password callback function PEM_cb() which either prompts for
16701 a password if usr_data is NULL or otherwise assumes it is a null
16702 terminated password. Allow passwords to be passed on command line
16703 environment or config files in a few more utilities.
16707 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16708 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16709 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16710 Update documentation.
16714 * Support for ASN1 "NULL" type. This could be handled before by using
16715 ASN1_TYPE but there wasn't any function that would try to read a NULL
16716 and produce an error if it couldn't. For compatibility we also have
16717 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16718 don't allocate anything because they don't need to.
16722 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16725 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16727 * Rebuild of the memory allocation routines used by OpenSSL code and
16728 possibly others as well. The purpose is to make an interface that
16729 provide hooks so anyone can build a separate set of allocation and
16730 deallocation routines to be used by OpenSSL, for example memory
16731 pool implementations, or something else, which was previously hard
16732 since Malloc(), Realloc() and Free() were defined as macros having
16733 the values malloc, realloc and free, respectively (except for Win32
16734 compilations). The same is provided for memory debugging code.
16735 OpenSSL already comes with functionality to find memory leaks, but
16736 this gives people a chance to debug other memory problems.
16738 With these changes, a new set of functions and macros have appeared:
16740 CRYPTO_set_mem_debug_functions() [F]
16741 CRYPTO_get_mem_debug_functions() [F]
16742 CRYPTO_dbg_set_options() [F]
16743 CRYPTO_dbg_get_options() [F]
16744 CRYPTO_malloc_debug_init() [M]
16746 The memory debug functions are NULL by default, unless the library
16747 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16748 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16749 gives the standard debugging functions that come with OpenSSL) or
16750 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16751 provided by the library user) must be used. When the standard
16752 debugging functions are used, CRYPTO_dbg_set_options can be used to
16753 request additional information:
16754 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16755 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16757 Also, things like CRYPTO_set_mem_functions will always give the
16758 expected result (the new set of functions is used for allocation
16759 and deallocation) at all times, regardless of platform and compiler
16762 To finish it up, some functions that were never use in any other
16763 way than through macros have a new API and new semantic:
16765 CRYPTO_dbg_malloc()
16766 CRYPTO_dbg_realloc()
16769 All macros of value have retained their old syntax.
16771 *Richard Levitte and Bodo Moeller*
16773 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16774 ordering of SMIMECapabilities wasn't in "strength order" and there
16775 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16780 * Some ASN1 types with illegal zero length encoding (INTEGER,
16781 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16783 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16785 * Merge in my S/MIME library for OpenSSL. This provides a simple
16786 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16787 functionality to handle multipart/signed properly) and a utility
16788 called 'smime' to call all this stuff. This is based on code I
16789 originally wrote for Celo who have kindly allowed it to be
16790 included in OpenSSL.
16794 * Add variants des_set_key_checked and des_set_key_unchecked of
16795 des_set_key (aka des_key_sched). Global variable des_check_key
16796 decides which of these is called by des_set_key; this way
16797 des_check_key behaves as it always did, but applications and
16798 the library itself, which was buggy for des_check_key == 1,
16799 have a cleaner way to pick the version they need.
16803 * New function PKCS12_newpass() which changes the password of a
16808 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16809 dynamic mix. In both cases the ids can be used as an index into the
16810 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16811 functions so they accept a list of the field values and the
16812 application doesn't need to directly manipulate the X509_TRUST
16817 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16822 * Modify the way the V3 extension code looks up extensions. This now
16823 works in a similar way to the object code: we have some "standard"
16824 extensions in a static table which is searched with OBJ_bsearch()
16825 and the application can add dynamic ones if needed. The file
16826 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16827 updated whenever a new extension is added to the core code and kept
16828 in ext_nid order. There is a simple program 'tabtest.c' which checks
16829 this. New extensions are not added too often so this file can readily
16830 be maintained manually.
16832 There are two big advantages in doing things this way. The extensions
16833 can be looked up immediately and no longer need to be "added" using
16834 X509V3_add_standard_extensions(): this function now does nothing.
16835 Side note: I get *lots* of email saying the extension code doesn't
16836 work because people forget to call this function.
16837 Also no dynamic allocation is done unless new extensions are added:
16838 so if we don't add custom extensions there is no need to call
16839 X509V3_EXT_cleanup().
16843 * Modify enc utility's salting as follows: make salting the default. Add a
16844 magic header, so unsalted files fail gracefully instead of just decrypting
16845 to garbage. This is because not salting is a big security hole, so people
16846 should be discouraged from doing it.
16850 * Fixes and enhancements to the 'x509' utility. It allowed a message
16851 digest to be passed on the command line but it only used this
16852 parameter when signing a certificate. Modified so all relevant
16853 operations are affected by the digest parameter including the
16854 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16855 DSA key was used because it didn't fix the digest.
16859 * Initial certificate chain verify code. Currently tests the untrusted
16860 certificates for consistency with the verify purpose (which is set
16861 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16863 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16864 this is because it will reject chains with invalid extensions whereas
16865 every previous version of OpenSSL and SSLeay made no checks at all.
16867 Trust code: checks the root CA for the relevant trust settings. Trust
16868 settings have an initial value consistent with the verify purpose: e.g.
16869 if the verify purpose is for SSL client use it expects the CA to be
16870 trusted for SSL client use. However the default value can be changed to
16871 permit custom trust settings: one example of this would be to only trust
16872 certificates from a specific "secure" set of CAs.
16874 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16875 which should be used for version portability: especially since the
16876 verify structure is likely to change more often now.
16878 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16879 to set them. If not set then assume SSL clients will verify SSL servers
16882 Two new options to the verify program: -untrusted allows a set of
16883 untrusted certificates to be passed in and -purpose which sets the
16884 intended purpose of the certificate. If a purpose is set then the
16885 new chain verify code is used to check extension consistency.
16889 * Support for the authority information access extension.
16893 * Modify RSA and DSA PEM read routines to transparently handle
16894 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16895 public keys in a format compatible with certificate
16896 SubjectPublicKeyInfo structures. Unfortunately there were already
16897 functions called *_PublicKey_* which used various odd formats so
16898 these are retained for compatibility: however the DSA variants were
16899 never in a public release so they have been deleted. Changed dsa/rsa
16900 utilities to handle the new format: note no releases ever handled public
16901 keys so we should be OK.
16903 The primary motivation for this change is to avoid the same fiasco
16904 that dogs private keys: there are several incompatible private key
16905 formats some of which are standard and some OpenSSL specific and
16906 require various evil hacks to allow partial transparent handling and
16907 even then it doesn't work with DER formats. Given the option anything
16908 other than PKCS#8 should be dumped: but the other formats have to
16909 stay in the name of compatibility.
16911 With public keys and the benefit of hindsight one standard format
16912 is used which works with EVP_PKEY, RSA or DSA structures: though
16913 it clearly returns an error if you try to read the wrong kind of key.
16915 Added a -pubkey option to the 'x509' utility to output the public key.
16916 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16917 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16918 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16919 that do the same as the `EVP_PKEY_assign_*()` except they up the
16920 reference count of the added key (they don't "swallow" the
16925 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16926 CRLs would fail if the file contained no certificates or no CRLs:
16927 added a new function to read in both types and return the number
16928 read: this means that if none are read it will be an error. The
16929 DER versions of the certificate and CRL reader would always fail
16930 because it isn't possible to mix certificates and CRLs in DER format
16931 without choking one or the other routine. Changed this to just read
16932 a certificate: this is the best we can do. Also modified the code
16933 in `apps/verify.c` to take notice of return codes: it was previously
16934 attempting to read in certificates from NULL pointers and ignoring
16935 any errors: this is one reason why the cert and CRL reader seemed
16936 to work. It doesn't check return codes from the default certificate
16937 routines: these may well fail if the certificates aren't installed.
16941 * Code to support otherName option in GeneralName.
16945 * First update to verify code. Change the verify utility
16946 so it warns if it is passed a self signed certificate:
16947 for consistency with the normal behaviour. X509_verify
16948 has been modified to it will now verify a self signed
16949 certificate if *exactly* the same certificate appears
16950 in the store: it was previously impossible to trust a
16951 single self signed certificate. This means that:
16952 openssl verify ss.pem
16953 now gives a warning about a self signed certificate but
16954 openssl verify -CAfile ss.pem ss.pem
16959 * For servers, store verify_result in SSL_SESSION data structure
16960 (and add it to external session representation).
16961 This is needed when client certificate verifications fails,
16962 but an application-provided verification callback (set by
16963 SSL_CTX_set_cert_verify_callback) allows accepting the session
16964 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16965 but returns 1): When the session is reused, we have to set
16966 ssl->verify_result to the appropriate error code to avoid
16969 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16971 * Fix a bug in the new PKCS#7 code: it didn't consider the
16972 case in PKCS7_dataInit() where the signed PKCS7 structure
16973 didn't contain any existing data because it was being created.
16975 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16977 * Add a salt to the key derivation routines in enc.c. This
16978 forms the first 8 bytes of the encrypted file. Also add a
16979 -S option to allow a salt to be input on the command line.
16983 * New function X509_cmp(). Oddly enough there wasn't a function
16984 to compare two certificates. We do this by working out the SHA1
16985 hash and comparing that. X509_cmp() will be needed by the trust
16990 * SSL_get1_session() is like SSL_get_session(), but increments
16991 the reference count in the SSL_SESSION returned.
16993 *Geoff Thorpe <geoff@eu.c2.net>*
16995 * Fix for 'req': it was adding a null to request attributes.
16996 Also change the X509_LOOKUP and X509_INFO code to handle
16997 certificate auxiliary information.
17001 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17006 * Add the possibility to add extra information to the memory leak
17007 detecting output, to form tracebacks, showing from where each
17008 allocation was originated: CRYPTO_push_info("constant string") adds
17009 the string plus current file name and line number to a per-thread
17010 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17011 is like calling CYRPTO_pop_info() until the stack is empty.
17012 Also updated memory leak detection code to be multi-thread-safe.
17016 * Add options -text and -noout to pkcs7 utility and delete the
17017 encryption options which never did anything. Update docs.
17021 * Add options to some of the utilities to allow the pass phrase
17022 to be included on either the command line (not recommended on
17023 OSes like Unix) or read from the environment. Update the
17024 manpages and fix a few bugs.
17028 * Add a few manpages for some of the openssl commands.
17032 * Fix the -revoke option in ca. It was freeing up memory twice,
17033 leaking and not finding already revoked certificates.
17037 * Extensive changes to support certificate auxiliary information.
17038 This involves the use of X509_CERT_AUX structure and X509_AUX
17039 functions. An X509_AUX function such as PEM_read_X509_AUX()
17040 can still read in a certificate file in the usual way but it
17041 will also read in any additional "auxiliary information". By
17042 doing things this way a fair degree of compatibility can be
17043 retained: existing certificates can have this information added
17044 using the new 'x509' options.
17046 Current auxiliary information includes an "alias" and some trust
17047 settings. The trust settings will ultimately be used in enhanced
17048 certificate chain verification routines: currently a certificate
17049 can only be trusted if it is self signed and then it is trusted
17054 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17055 The problem was that one of the replacement routines had not been working
17056 since SSLeay releases. For now the offending routine has been replaced
17057 with non-optimised assembler. Even so, this now gives around 95%
17058 performance improvement for 1024 bit RSA signs.
17062 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17063 handling. Most clients have the effective key size in bits equal to
17064 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17065 A few however don't do this and instead use the size of the decrypted key
17066 to determine the RC2 key length and the AlgorithmIdentifier to determine
17067 the effective key length. In this case the effective key length can still
17068 be 40 bits but the key length can be 168 bits for example. This is fixed
17069 by manually forcing an RC2 key into the EVP_PKEY structure because the
17070 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17071 the key length and effective key length are equal.
17075 * Add a bunch of functions that should simplify the creation of
17076 X509_NAME structures. Now you should be able to do:
17077 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17078 and have it automatically work out the correct field type and fill in
17079 the structures. The more adventurous can try:
17080 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17081 and it will (hopefully) work out the correct multibyte encoding.
17085 * Change the 'req' utility to use the new field handling and multibyte
17086 copy routines. Before the DN field creation was handled in an ad hoc
17087 way in req, ca, and x509 which was rather broken and didn't support
17088 BMPStrings or UTF8Strings. Since some software doesn't implement
17089 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17090 using the dirstring_type option. See the new comment in the default
17091 openssl.cnf for more info.
17095 * Make crypto/rand/md_rand.c more robust:
17096 - Assure unique random numbers after fork().
17097 - Make sure that concurrent threads access the global counter and
17098 md serializably so that we never lose entropy in them
17099 or use exactly the same state in multiple threads.
17100 Access to the large state is not always serializable because
17101 the additional locking could be a performance killer, and
17102 md should be large enough anyway.
17106 * New file `apps/app_rand.c` with commonly needed functionality
17107 for handling the random seed file.
17109 Use the random seed file in some applications that previously did not:
17111 dsaparam -genkey (which also ignored its '-rand' option),
17114 x509 (when signing).
17115 Except on systems with /dev/urandom, it is crucial to have a random
17116 seed file at least for key creation, DSA signing, and for DH exchanges;
17117 for RSA signatures we could do without one.
17119 gendh and gendsa (unlike genrsa) used to read only the first byte
17120 of each file listed in the '-rand' option. The function as previously
17121 found in genrsa is now in app_rand.c and is used by all programs
17122 that support '-rand'.
17126 * In RAND_write_file, use mode 0600 for creating files;
17127 don't just chmod when it may be too late.
17131 * Report an error from X509_STORE_load_locations
17132 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17136 * New function ASN1_mbstring_copy() this copies a string in either
17137 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17138 into an ASN1_STRING type. A mask of permissible types is passed
17139 and it chooses the "minimal" type to use or an error if not type
17144 * Add function equivalents to the various macros in asn1.h. The old
17145 macros are retained with an `M_` prefix. Code inside the library can
17146 use the `M_` macros. External code (including the openssl utility)
17147 should *NOT* in order to be "shared library friendly".
17151 * Add various functions that can check a certificate's extensions
17152 to see if it usable for various purposes such as SSL client,
17153 server or S/MIME and CAs of these types. This is currently
17154 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17155 verification. Also added a -purpose flag to x509 utility to
17156 print out all the purposes.
17160 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17165 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17166 for, obtain and decode and extension and obtain its critical flag.
17167 This allows all the necessary extension code to be handled in a
17168 single function call.
17172 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17173 platforms. See crypto/rc4/rc4_enc.c for further details.
17177 * New -noout option to asn1parse. This causes no output to be produced
17178 its main use is when combined with -strparse and -out to extract data
17179 from a file (which may not be in ASN.1 format).
17183 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17184 when producing the local key id.
17186 *Richard Levitte <levitte@stacken.kth.se>*
17188 * New option -dhparam in s_server. This allows a DH parameter file to be
17189 stated explicitly. If it is not stated then it tries the first server
17190 certificate file. The previous behaviour hard coded the filename
17195 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17196 a public key to be input or output. For example:
17197 openssl rsa -in key.pem -pubout -out pubkey.pem
17198 Also added necessary DSA public key functions to handle this.
17202 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17203 in the message. This was handled by allowing
17204 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17206 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17208 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17209 to the end of the strings whereas this didn't. This would cause problems
17210 if strings read with d2i_ASN1_bytes() were later modified.
17212 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17214 * Fix for base64 decode bug. When a base64 bio reads only one line of
17215 data and it contains EOF it will end up returning an error. This is
17216 caused by input 46 bytes long. The cause is due to the way base64
17217 BIOs find the start of base64 encoded data. They do this by trying a
17218 trial decode on each line until they find one that works. When they
17219 do a flag is set and it starts again knowing it can pass all the
17220 data directly through the decoder. Unfortunately it doesn't reset
17221 the context it uses. This means that if EOF is reached an attempt
17222 is made to pass two EOFs through the context and this causes the
17223 resulting error. This can also cause other problems as well. As is
17224 usual with these problems it takes *ages* to find and the fix is
17225 trivial: move one line.
17227 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17229 * Ugly workaround to get s_client and s_server working under Windows. The
17230 old code wouldn't work because it needed to select() on sockets and the
17231 tty (for keypresses and to see if data could be written). Win32 only
17232 supports select() on sockets so we select() with a 1s timeout on the
17233 sockets and then see if any characters are waiting to be read, if none
17234 are present then we retry, we also assume we can always write data to
17235 the tty. This isn't nice because the code then blocks until we've
17236 received a complete line of data and it is effectively polling the
17237 keyboard at 1s intervals: however it's quite a bit better than not
17238 working at all :-) A dedicated Windows application might handle this
17239 with an event loop for example.
17243 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17244 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17245 will be called when RSA_sign() and RSA_verify() are used. This is useful
17246 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17247 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17248 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17249 This necessitated the support of an extra signature type NID_md5_sha1
17250 for SSL signatures and modifications to the SSL library to use it instead
17251 of calling RSA_public_decrypt() and RSA_private_encrypt().
17255 * Add new -verify -CAfile and -CApath options to the crl program, these
17256 will lookup a CRL issuers certificate and verify the signature in a
17257 similar way to the verify program. Tidy up the crl program so it
17258 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17259 less strict. It will now permit CRL extensions even if it is not
17260 a V2 CRL: this will allow it to tolerate some broken CRLs.
17264 * Initialize all non-automatic variables each time one of the openssl
17265 sub-programs is started (this is necessary as they may be started
17266 multiple times from the "OpenSSL>" prompt).
17268 *Lennart Bang, Bodo Moeller*
17270 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17271 removing all other RSA functionality (this is what NO_RSA does). This
17272 is so (for example) those in the US can disable those operations covered
17273 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17278 * Non-copying interface to BIO pairs.
17279 (still largely untested)
17283 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17284 ASCII string. This was handled independently in various places before.
17288 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17289 UTF8 strings a character at a time.
17293 * Use client_version from client hello to select the protocol
17294 (s23_srvr.c) and for RSA client key exchange verification
17295 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17299 * Add various utility functions to handle SPKACs, these were previously
17300 handled by poking round in the structure internals. Added new function
17301 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17302 print, verify and generate SPKACs. Based on an original idea from
17303 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17307 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17311 * Allow the config file extension section to be overwritten on the
17312 command line. Based on an original idea from Massimiliano Pala
17313 <madwolf@comune.modena.it>. The new option is called -extensions
17314 and can be applied to ca, req and x509. Also -reqexts to override
17315 the request extensions in req and -crlexts to override the crl extensions
17320 * Add new feature to the SPKAC handling in ca. Now you can include
17321 the same field multiple times by preceding it by "XXXX." for example:
17324 this is the same syntax as used in the req config file.
17328 * Allow certificate extensions to be added to certificate requests. These
17329 are specified in a 'req_extensions' option of the req section of the
17330 config file. They can be printed out with the -text option to req but
17331 are otherwise ignored at present.
17335 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17336 data read consists of only the final block it would not decrypted because
17337 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17338 A misplaced 'break' also meant the decrypted final block might not be
17339 copied until the next read.
17343 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17344 a few extra parameters to the DH structure: these will be useful if
17345 for example we want the value of 'q' or implement X9.42 DH.
17349 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17350 provides hooks that allow the default DSA functions or functions on a
17351 "per key" basis to be replaced. This allows hardware acceleration and
17352 hardware key storage to be handled without major modification to the
17353 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17354 associated functions.
17358 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17359 as "read only": it can't be written to and the buffer it points to will
17360 not be freed. Reading from a read only BIO is much more efficient than
17361 a normal memory BIO. This was added because there are several times when
17362 an area of memory needs to be read from a BIO. The previous method was
17363 to create a memory BIO and write the data to it, this results in two
17364 copies of the data and an O(n^2) reading algorithm. There is a new
17365 function BIO_new_mem_buf() which creates a read only memory BIO from
17366 an area of memory. Also modified the PKCS#7 routines to use read only
17371 * Bugfix: ssl23_get_client_hello did not work properly when called in
17372 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17373 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17374 but a retry condition occurred while trying to read the rest.
17378 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17379 NID_pkcs7_encrypted by default: this was wrong since this should almost
17380 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17381 the encrypted data type: this is a more sensible place to put it and it
17382 allows the PKCS#12 code to be tidied up that duplicated this
17387 * Changed obj_dat.pl script so it takes its input and output files on
17388 the command line. This should avoid shell escape redirection problems
17393 * Initial support for certificate extension requests, these are included
17394 in things like Xenroll certificate requests. Included functions to allow
17395 extensions to be obtained and added.
17399 * -crlf option to s_client and s_server for sending newlines as
17400 CRLF (as required by many protocols).
17404 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17406 * Install libRSAglue.a when OpenSSL is built with RSAref.
17408 *Ralf S. Engelschall*
17410 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17412 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17414 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17419 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17420 DH parameters/keys (q is lost during that conversion, but the resulting
17421 DH parameters contain its length).
17423 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17424 much faster than DH_generate_parameters (which creates parameters
17425 where `p = 2*q + 1`), and also the smaller q makes DH computations
17426 much more efficient (160-bit exponentiation instead of 1024-bit
17427 exponentiation); so this provides a convenient way to support DHE
17428 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17429 utter importance to use
17430 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17432 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17433 when such DH parameters are used, because otherwise small subgroup
17434 attacks may become possible!
17438 * Avoid memory leak in i2d_DHparams.
17442 * Allow the -k option to be used more than once in the enc program:
17443 this allows the same encrypted message to be read by multiple recipients.
17447 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17448 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17449 it will always use the numerical form of the OID, even if it has a short
17454 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17455 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17456 otherwise bn_mod_exp was called. In the case of hardware keys for example
17457 no private key components need be present and it might store extra data
17458 in the RSA structure, which cannot be accessed from bn_mod_exp.
17459 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17460 private key operations.
17464 * Added support for SPARC Linux.
17468 * pem_password_cb function type incompatibly changed from
17469 typedef int pem_password_cb(char *buf, int size, int rwflag);
17471 ....(char *buf, int size, int rwflag, void *userdata);
17472 so that applications can pass data to their callbacks:
17473 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17474 additional void * argument, which is just handed through whenever
17475 the password callback is called.
17477 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17479 New function SSL_CTX_set_default_passwd_cb_userdata.
17481 Compatibility note: As many C implementations push function arguments
17482 onto the stack in reverse order, the new library version is likely to
17483 interoperate with programs that have been compiled with the old
17484 pem_password_cb definition (PEM_whatever takes some data that
17485 happens to be on the stack as its last argument, and the callback
17486 just ignores this garbage); but there is no guarantee whatsoever that
17489 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17490 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17491 problems not only on Windows, but also on some Unix platforms.
17492 To avoid problematic command lines, these definitions are now in an
17493 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17494 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17498 * MIPS III/IV assembler module is reimplemented.
17502 * More DES library cleanups: remove references to srand/rand and
17503 delete an unused file.
17507 * Add support for the free Netwide assembler (NASM) under Win32,
17508 since not many people have MASM (ml) and it can be hard to obtain.
17509 This is currently experimental but it seems to work OK and pass all
17510 the tests. Check out INSTALL.W32 for info.
17514 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17515 without temporary keys kept an extra copy of the server key,
17516 and connections with temporary keys did not free everything in case
17521 * New function RSA_check_key and new openssl rsa option -check
17522 for verifying the consistency of RSA keys.
17524 *Ulf Moeller, Bodo Moeller*
17526 * Various changes to make Win32 compile work:
17527 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17528 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17529 comparison" warnings.
17530 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17534 * Add a debugging option to PKCS#5 v2 key generation function: when
17535 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17536 derived keys are printed to stderr.
17540 * Copy the flags in ASN1_STRING_dup().
17542 *Roman E. Pavlov <pre@mo.msk.ru>*
17544 * The x509 application mishandled signing requests containing DSA
17545 keys when the signing key was also DSA and the parameters didn't match.
17547 It was supposed to omit the parameters when they matched the signing key:
17548 the verifying software was then supposed to automatically use the CA's
17549 parameters if they were absent from the end user certificate.
17551 Omitting parameters is no longer recommended. The test was also
17552 the wrong way round! This was probably due to unusual behaviour in
17553 EVP_cmp_parameters() which returns 1 if the parameters match.
17554 This meant that parameters were omitted when they *didn't* match and
17555 the certificate was useless. Certificates signed with 'ca' didn't have
17558 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17560 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17561 The interface is as follows:
17562 Applications can use
17563 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17564 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17565 "off" is now the default.
17566 The library internally uses
17567 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17568 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17569 to disable memory-checking temporarily.
17571 Some inconsistent states that previously were possible (and were
17572 even the default) are now avoided.
17574 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17575 with each memory chunk allocated; this is occasionally more helpful
17576 than just having a counter.
17578 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17580 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17585 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17586 which largely parallels "options", but is for changing API behaviour,
17587 whereas "options" are about protocol behaviour.
17588 Initial "mode" flags are:
17590 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17591 a single record has been written.
17592 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17593 retries use the same buffer location.
17594 (But all of the contents must be
17599 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17602 * Fix problems with no-hmac etc.
17604 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17606 * New functions RSA_get_default_method(), RSA_set_method() and
17607 RSA_get_method(). These allows replacement of RSA_METHODs without having
17608 to mess around with the internals of an RSA structure.
17612 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17613 Also really enable memory leak checks in openssl.c and in some
17616 *Chad C. Mulligan, Bodo Moeller*
17618 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17619 up the length of negative integers. This has now been simplified to just
17620 store the length when it is first determined and use it later, rather
17621 than trying to keep track of where data is copied and updating it to
17623 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17625 * Add a new function PKCS7_signatureVerify. This allows the verification
17626 of a PKCS#7 signature but with the signing certificate passed to the
17627 function itself. This contrasts with PKCS7_dataVerify which assumes the
17628 certificate is present in the PKCS#7 structure. This isn't always the
17629 case: certificates can be omitted from a PKCS#7 structure and be
17630 distributed by "out of band" means (such as a certificate database).
17634 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17635 function prototypes in pem.h, also change util/mkdef.pl to add the
17636 necessary function names.
17640 * mk1mf.pl (used by Windows builds) did not properly read the
17641 options set by Configure in the top level Makefile, and Configure
17642 was not even able to write more than one option correctly.
17643 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17647 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17648 file to be loaded from a BIO or FILE pointer. The BIO version will
17649 for example allow memory BIOs to contain config info.
17653 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17654 Whoever hopes to achieve shared-library compatibility across versions
17655 must use this, not the compile-time macro.
17656 (Exercise 0.9.4: Which is the minimum library version required by
17658 Note: All this applies only to multi-threaded programs, others don't
17663 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17664 through a BIO pair triggered the default case, i.e.
17665 SSLerr(...,SSL_R_UNKNOWN_STATE).
17669 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17670 can use the SSL library even if none of the specific BIOs is
17675 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17676 for the encoded length.
17678 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17680 * Add initial documentation of the X509V3 functions.
17684 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17685 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17686 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17687 secure PKCS#8 private key format with a high iteration count.
17691 * Fix determination of Perl interpreter: A perl or perl5
17692 *directory* in $PATH was also accepted as the interpreter.
17694 *Ralf S. Engelschall*
17696 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17697 wrong with it but it was very old and did things like calling
17698 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17699 unusual formatting.
17703 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17704 to use the new extension code.
17708 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17709 with macros. This should make it easier to change their form, add extra
17710 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17715 * Add to configuration table a new entry that can specify an alternative
17716 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17717 according to Mark Crispin <MRC@Panda.COM>.
17721 * DES CBC did not update the IV. Weird.
17725 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17726 Changing the behaviour of the former might break existing programs --
17727 where IV updating is needed, des_ncbc_encrypt can be used.
17730 * When bntest is run from "make test" it drives bc to check its
17731 calculations, as well as internally checking them. If an internal check
17732 fails, it needs to cause bc to give a non-zero result or make test carries
17733 on without noticing the failure. Fixed.
17737 * DES library cleanups.
17741 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17742 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17743 ciphers. NOTE: although the key derivation function has been verified
17744 against some published test vectors it has not been extensively tested
17745 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17750 * Instead of "mkdir -p", which is not fully portable, use new
17751 Perl script "util/mkdir-p.pl".
17755 * Rewrite the way password based encryption (PBE) is handled. It used to
17756 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17757 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17758 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17759 the 'parameter' field of the AlgorithmIdentifier is passed to the
17760 underlying key generation function so it must do its own ASN1 parsing.
17761 This has also changed the EVP_PBE_CipherInit() function which now has a
17762 'parameter' argument instead of literal salt and iteration count values
17763 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17767 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17768 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17769 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17770 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17771 value was just used as a "magic string" and not used directly its
17772 value doesn't matter.
17776 * Introduce some semblance of const correctness to BN. Shame C doesn't
17781 * "linux-sparc64" configuration (ultrapenguin).
17783 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17784 "linux-sparc" configuration.
17786 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17788 * config now generates no-xxx options for missing ciphers.
17792 * Support the EBCDIC character set (work in progress).
17793 File ebcdic.c not yet included because it has a different license.
17795 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17797 * Support BS2000/OSD-POSIX.
17799 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17801 * Make callbacks for key generation use `void *` instead of `char *`.
17805 * Make S/MIME samples compile (not yet tested).
17809 * Additional typesafe stacks.
17813 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17817 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17819 * New configuration variant "sco5-gcc".
17821 * Updated some demos.
17823 *Sean O Riordain, Wade Scholine*
17825 * Add missing BIO_free at exit of pkcs12 application.
17829 * Fix memory leak in conf.c.
17833 * Updates for Win32 to assembler version of MD5.
17837 * Set #! path to perl in `apps/der_chop` to where we found it
17838 instead of using a fixed path.
17842 * SHA library changes for irix64-mips4-cc.
17846 * Improvements for VMS support.
17850 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17852 * Bignum library bug fix. IRIX 6 passes "make test" now!
17853 This also avoids the problems with SC4.2 and unpatched SC5.
17855 *Andy Polyakov <appro@fy.chalmers.se>*
17857 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17858 These are required because of the typesafe stack would otherwise break
17859 existing code. If old code used a structure member which used to be STACK
17860 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17861 sk_num or sk_value it would produce an error because the num, data members
17862 are not present in STACK_OF. Now it just produces a warning. sk_set
17863 replaces the old method of assigning a value to sk_value
17864 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17865 that does this will no longer work (and should use sk_set instead) but
17866 this could be regarded as a "questionable" behaviour anyway.
17870 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17871 correctly handle encrypted S/MIME data.
17875 * Change type of various DES function arguments from des_cblock
17876 (which means, in function argument declarations, pointer to char)
17877 to des_cblock * (meaning pointer to array with 8 char elements),
17878 which allows the compiler to do more typechecking; it was like
17879 that back in SSLeay, but with lots of ugly casts.
17881 Introduce new type const_des_cblock.
17885 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17886 problems: find RecipientInfo structure that matches recipient certificate
17887 and initialise the ASN1 structures properly based on passed cipher.
17891 * Belatedly make the BN tests actually check the results.
17895 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17896 to and from BNs: it was completely broken. New compilation option
17897 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17898 key elements as negative integers.
17902 * Reorganize and speed up MD5.
17904 *Andy Polyakov <appro@fy.chalmers.se>*
17908 *Richard Levitte <richard@levitte.org>*
17910 * New option -out to asn1parse to allow the parsed structure to be
17911 output to a file. This is most useful when combined with the -strparse
17912 option to examine the output of things like OCTET STRINGS.
17916 * Make SSL library a little more fool-proof by not requiring any longer
17917 that `SSL_set_{accept,connect}_state` be called before
17918 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17919 in many applications because usually everything *appeared* to work as
17920 intended anyway -- now it really works as intended).
17924 * Move openssl.cnf out of lib/.
17928 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17929 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17930 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17932 *Ralf S. Engelschall*
17934 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17935 handle PKCS#7 enveloped data properly.
17937 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17939 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17940 copying pointers. The cert_st handling is changed by this in
17941 various ways (and thus what used to be known as ctx->default_cert
17942 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17943 any longer when s->cert does not give us what we need).
17944 ssl_cert_instantiate becomes obsolete by this change.
17945 As soon as we've got the new code right (possibly it already is?),
17946 we have solved a couple of bugs of the earlier code where s->cert
17947 was used as if it could not have been shared with other SSL structures.
17949 Note that using the SSL API in certain dirty ways now will result
17950 in different behaviour than observed with earlier library versions:
17951 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17952 does not influence s as it used to.
17954 In order to clean up things more thoroughly, inside SSL_SESSION
17955 we don't use CERT any longer, but a new structure SESS_CERT
17956 that holds per-session data (if available); currently, this is
17957 the peer's certificate chain and, for clients, the server's certificate
17958 and temporary key. CERT holds only those values that can have
17959 meaningful defaults in an SSL_CTX.
17963 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17964 from the internal representation. Various PKCS#7 fixes: remove some
17965 evil casts and set the enc_dig_alg field properly based on the signing
17970 * Allow PKCS#12 password to be set from the command line or the
17971 environment. Let 'ca' get its config file name from the environment
17972 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17977 * Allow certificate policies extension to use an IA5STRING for the
17978 organization field. This is contrary to the PKIX definition but
17979 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17984 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17985 without disallowing inline assembler and the like for non-pedantic builds.
17989 * Support Borland C++ builder.
17991 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17997 * SHA-1 cleanups and performance enhancements.
17999 *Andy Polyakov <appro@fy.chalmers.se>*
18001 * Sparc v8plus assembler for the bignum library.
18003 *Andy Polyakov <appro@fy.chalmers.se>*
18005 * Accept any -xxx and +xxx compiler options in Configure.
18009 * Update HPUX configuration.
18013 * Add missing `sk_<type>_unshift()` function to safestack.h
18015 *Ralf S. Engelschall*
18017 * New function SSL_CTX_use_certificate_chain_file that sets the
18018 "extra_cert"s in addition to the certificate. (This makes sense
18019 only for "PEM" format files, as chains as a whole are not
18024 * Support verify_depth from the SSL API.
18025 x509_vfy.c had what can be considered an off-by-one-error:
18026 Its depth (which was not part of the external interface)
18027 was actually counting the number of certificates in a chain;
18028 now it really counts the depth.
18032 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18033 instead of X509err, which often resulted in confusing error
18034 messages since the error codes are not globally unique
18035 (e.g. an alleged error in ssl3_accept when a certificate
18036 didn't match the private key).
18038 * New function SSL_CTX_set_session_id_context that allows to set a default
18039 value (so that you don't need SSL_set_session_id_context for each
18040 connection using the SSL_CTX).
18044 * OAEP decoding bug fix.
18048 * Support INSTALL_PREFIX for package builders, as proposed by
18053 * New Configure options "threads" and "no-threads". For systems
18054 where the proper compiler options are known (currently Solaris
18055 and Linux), "threads" is the default.
18059 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18063 * Install various scripts to $(OPENSSLDIR)/misc, not to
18064 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18065 such as /usr/local/bin.
18069 * "make linux-shared" to build shared libraries.
18071 *Niels Poppe <niels@netbox.org>*
18073 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18077 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18078 extension adding in x509 utility.
18082 * Remove NOPROTO sections and error code comments.
18086 * Partial rewrite of the DEF file generator to now parse the ANSI
18091 * New Configure options --prefix=DIR and --openssldir=DIR.
18095 * Complete rewrite of the error code script(s). It is all now handled
18096 by one script at the top level which handles error code gathering,
18097 header rewriting and C source file generation. It should be much better
18098 than the old method: it now uses a modified version of Ulf's parser to
18099 read the ANSI prototypes in all header files (thus the old K&R definitions
18100 aren't needed for error creation any more) and do a better job of
18101 translating function codes into names. The old 'ASN1 error code embedded
18102 in a comment' is no longer necessary and it doesn't use .err files which
18103 have now been deleted. Also the error code call doesn't have to appear all
18104 on one line (which resulted in some large lines...).
18108 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18112 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18113 0 (which usually indicates a closed connection), but continue reading.
18117 * Fix some race conditions.
18121 * Add support for CRL distribution points extension. Add Certificate
18122 Policies and CRL distribution points documentation.
18126 * Move the autogenerated header file parts to crypto/opensslconf.h.
18130 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18131 8 of keying material. Merlin has also confirmed interop with this fix
18132 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18134 *Merlin Hughes <merlin@baltimore.ie>*
18136 * Fix lots of warnings.
18138 *Richard Levitte <levitte@stacken.kth.se>*
18140 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18141 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18143 *Richard Levitte <levitte@stacken.kth.se>*
18145 * Fix problems with sizeof(long) == 8.
18147 *Andy Polyakov <appro@fy.chalmers.se>*
18149 * Change functions to ANSI C.
18153 * Fix typos in error codes.
18155 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18157 * Remove defunct assembler files from Configure.
18161 * SPARC v8 assembler BIGNUM implementation.
18163 *Andy Polyakov <appro@fy.chalmers.se>*
18165 * Support for Certificate Policies extension: both print and set.
18166 Various additions to support the r2i method this uses.
18170 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18171 return a const string when you are expecting an allocated buffer.
18175 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18176 types DirectoryString and DisplayText.
18180 * Add code to allow r2i extensions to access the configuration database,
18181 add an LHASH database driver and add several ctx helper functions.
18185 * Fix an evil bug in bn_expand2() which caused various BN functions to
18186 fail when they extended the size of a BIGNUM.
18190 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18191 support typesafe stack.
18195 * Fix typo in SSL_[gs]et_options().
18197 *Nils Frostberg <nils@medcom.se>*
18199 * Delete various functions and files that belonged to the (now obsolete)
18200 old X509V3 handling code.
18204 * New Configure option "rsaref".
18208 * Don't auto-generate pem.h.
18212 * Introduce type-safe ASN.1 SETs.
18216 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18218 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18220 * Introduce type-safe STACKs. This will almost certainly break lots of code
18221 that links with OpenSSL (well at least cause lots of warnings), but fear
18222 not: the conversion is trivial, and it eliminates loads of evil casts. A
18223 few STACKed things have been converted already. Feel free to convert more.
18224 In the fullness of time, I'll do away with the STACK type altogether.
18228 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18229 specified in `<certfile>` by updating the entry in the index.txt file.
18230 This way one no longer has to edit the index.txt file manually for
18231 revoking a certificate. The -revoke option does the gory details now.
18233 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18235 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18236 `-text` option at all and this way the `-noout -text` combination was
18237 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18239 *Ralf S. Engelschall*
18241 * Make sure a corresponding plain text error message exists for the
18242 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18243 verify callback function determined that a certificate was revoked.
18245 *Ralf S. Engelschall*
18247 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18248 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18249 all available ciphers including rc5, which was forgotten until now.
18250 In order to let the testing shell script know which algorithms
18251 are available, a new (up to now undocumented) command
18252 `openssl list-cipher-commands` is used.
18256 * Bugfix: s_client occasionally would sleep in select() when
18257 it should have checked SSL_pending() first.
18261 * New functions DSA_do_sign and DSA_do_verify to provide access to
18262 the raw DSA values prior to ASN.1 encoding.
18266 * Tweaks to Configure
18268 *Niels Poppe <niels@netbox.org>*
18270 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18275 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18279 * New config option to avoid instructions that are illegal on the 80386.
18280 The default code is faster, but requires at least a 486.
18284 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18285 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18286 same as SSL2_VERSION anyway.
18290 * New "-showcerts" option for s_client.
18294 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18295 application. Various cleanups and fixes.
18299 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18300 modify error routines to work internally. Add error codes and PBE init
18301 to library startup routines.
18305 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18306 packing functions to asn1 and evp. Changed function names and error
18307 codes along the way.
18311 * PKCS12 integration: and so it begins... First of several patches to
18312 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18313 objects to objects.h
18317 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18318 and display support for Thawte strong extranet extension.
18322 * Add LinuxPPC support.
18324 *Jeff Dubrule <igor@pobox.org>*
18326 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18327 bn_div_words in alpha.s.
18329 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18331 * Make sure the RSA OAEP test is skipped under -DRSAref because
18332 OAEP isn't supported when OpenSSL is built with RSAref.
18334 *Ulf Moeller <ulf@fitug.de>*
18336 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18337 so they no longer are missing under -DNOPROTO.
18339 *Soren S. Jorvang <soren@t.dk>*
18341 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18343 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18344 doesn't work when the session is reused. Coming soon!
18348 * Fix a security hole, that allows sessions to be reused in the wrong
18349 context thus bypassing client cert protection! All software that uses
18350 client certs and session caches in multiple contexts NEEDS PATCHING to
18351 allow session reuse! A fuller solution is in the works.
18353 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18355 * Some more source tree cleanups (removed obsolete files
18356 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18357 permission on "config" script to be executable) and a fix for the INSTALL
18360 *Ulf Moeller <ulf@fitug.de>*
18362 * Remove some legacy and erroneous uses of malloc, free instead of
18365 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18367 * Make rsa_oaep_test return non-zero on error.
18369 *Ulf Moeller <ulf@fitug.de>*
18371 * Add support for native Solaris shared libraries. Configure
18372 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18373 if someone would make that last step automatic.
18375 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18377 * ctx_size was not built with the right compiler during "make links". Fixed.
18381 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18382 except NULL ciphers". This means the default cipher list will no longer
18383 enable NULL ciphers. They need to be specifically enabled e.g. with
18384 the string "DEFAULT:eNULL".
18388 * Fix to RSA private encryption routines: if p < q then it would
18389 occasionally produce an invalid result. This will only happen with
18390 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18394 * Be less restrictive and allow also `perl util/perlpath.pl
18395 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18396 because this way one can also use an interpreter named `perl5` (which is
18397 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18398 installed as `perl`).
18400 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18402 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18404 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18406 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18407 advapi32.lib to Win32 build and change the pem test comparison
18408 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18409 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18410 and crypto/des/ede_cbcm_enc.c.
18414 * DES quad checksum was broken on big-endian architectures. Fixed.
18418 * Comment out two functions in bio.h that aren't implemented. Fix up the
18419 Win32 test batch file so it (might) work again. The Win32 test batch file
18420 is horrible: I feel ill....
18424 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18425 in e_os.h. Audit of header files to check ANSI and non ANSI
18426 sections: 10 functions were absent from non ANSI section and not exported
18427 from Windows DLLs. Fixed up libeay.num for new functions.
18431 * Make `openssl version` output lines consistent.
18433 *Ralf S. Engelschall*
18435 * Fix Win32 symbol export lists for BIO functions: Added
18436 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18437 to ms/libeay{16,32}.def.
18439 *Ralf S. Engelschall*
18441 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18442 fine under Unix and passes some trivial tests I've now added. But the
18443 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18444 added to make sure no one expects that this stuff really works in the
18445 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18446 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18449 *Ralf S. Engelschall*
18451 * Fix the generation of two part addresses in perl.
18453 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18455 * Add config entry for Linux on MIPS.
18457 *John Tobey <jtobey@channel1.com>*
18459 * Make links whenever Configure is run, unless we are on Windoze.
18463 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18464 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18469 * Add a useful kludge to allow package maintainers to specify compiler and
18470 other platforms details on the command line without having to patch the
18471 Configure script every time: One now can use
18472 `perl Configure <id>:<details>`,
18473 i.e. platform ids are allowed to have details appended
18474 to them (separated by colons). This is treated as there would be a static
18475 pre-configured entry in Configure's %table under key `<id>` with value
18476 `<details>` and `perl Configure <id>` is called. So, when you want to
18477 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18478 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18479 now, which overrides the FreeBSD-elf entry on-the-fly.
18481 *Ralf S. Engelschall*
18483 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18487 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18488 on the `perl Configure ...` command line. This way one can compile
18489 OpenSSL libraries with Position Independent Code (PIC) which is needed
18490 for linking it into DSOs.
18492 *Ralf S. Engelschall*
18494 * Remarkably, export ciphers were totally broken and no-one had noticed!
18499 * Cleaned up the LICENSE document: The official contact for any license
18500 questions now is the OpenSSL core team under openssl-core@openssl.org.
18501 And add a paragraph about the dual-license situation to make sure people
18502 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18503 to the OpenSSL toolkit.
18505 *Ralf S. Engelschall*
18507 * General source tree makefile cleanups: Made `making xxx in yyy...`
18508 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18509 Additionally cleaned up the `make links` target: Remove unnecessary
18510 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18511 to speed processing and no longer clutter the display with confusing
18512 stuff. Instead only the actually done links are displayed.
18514 *Ralf S. Engelschall*
18516 * Permit null encryption ciphersuites, used for authentication only. It used
18517 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18518 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18523 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18524 signed attributes when verifying signatures (this would break them),
18525 the detached data encoding was wrong and public keys obtained using
18526 X509_get_pubkey() weren't freed.
18530 * Add text documentation for the BUFFER functions. Also added a work around
18531 to a Win95 console bug. This was triggered by the password read stuff: the
18532 last character typed gets carried over to the next fread(). If you were
18533 generating a new cert request using 'req' for example then the last
18534 character of the passphrase would be CR which would then enter the first
18539 * Added the new 'Includes OpenSSL Cryptography Software' button as
18540 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18541 button and can be used by applications based on OpenSSL to show the
18542 relationship to the OpenSSL project.
18544 *Ralf S. Engelschall*
18546 * Remove confusing variables in function signatures in files
18547 ssl/ssl_lib.c and ssl/ssl.h.
18549 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18551 * Don't install bss_file.c under PREFIX/include/
18553 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18555 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18556 functions that return function pointers and has support for NT specific
18557 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18558 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18559 unsigned to signed types: this was killing the Win32 compile.
18563 * Add new certificate file to stack functions,
18564 SSL_add_dir_cert_subjects_to_stack() and
18565 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18566 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18567 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18568 This means that Apache-SSL and similar packages don't have to mess around
18569 to add as many CAs as they want to the preferred list.
18573 * Experiment with doxygen documentation. Currently only partially applied to
18575 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18576 openssl.doxy as the configuration file.
18580 * Get rid of remaining C++-style comments which strict C compilers hate.
18582 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18584 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18585 compiled in by default: it has problems with large keys.
18589 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18590 DH private keys and/or callback functions which directly correspond to
18591 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18592 is needed for applications which have to configure certificates on a
18593 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18595 For the RSA certificate situation is makes no difference, but
18596 for the DSA certificate situation this fixes the "no shared cipher"
18597 problem where the OpenSSL cipher selection procedure failed because the
18598 temporary keys were not overtaken from the context and the API provided
18599 no way to reconfigure them.
18600 The new functions now let applications reconfigure the stuff and they
18601 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18602 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18603 non-public-API function ssl_cert_instantiate() is used as a helper
18604 function and also to reduce code redundancy inside ssl_rsa.c.
18606 *Ralf S. Engelschall*
18608 * Move s_server -dcert and -dkey options out of the undocumented feature
18609 area because they are useful for the DSA situation and should be
18610 recognized by the users.
18612 *Ralf S. Engelschall*
18614 * Fix the cipher decision scheme for export ciphers: the export bits are
18615 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18616 SSL_EXP_MASK. So, the original variable has to be used instead of the
18617 already masked variable.
18619 *Richard Levitte <levitte@stacken.kth.se>*
18621 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18623 *Richard Levitte <levitte@stacken.kth.se>*
18625 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18626 from `int` to `unsigned int` because it is a length and initialized by
18627 EVP_DigestFinal() which expects an `unsigned int *`.
18629 *Richard Levitte <levitte@stacken.kth.se>*
18631 * Don't hard-code path to Perl interpreter on shebang line of Configure
18632 script. Instead use the usual Shell->Perl transition trick.
18634 *Ralf S. Engelschall*
18636 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18637 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18638 -noout -modulus` as it's already the case for `openssl rsa -noout
18639 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18640 currently the public key is printed (a decision which was already done by
18641 `openssl dsa -modulus` in the past) which serves a similar purpose.
18642 Additionally the NO_RSA no longer completely removes the whole -modulus
18643 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18646 *Ralf S. Engelschall*
18648 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18649 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18651 *Arne Ansper <arne@ats.cyber.ee>*
18653 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18654 to be added. Now both 'req' and 'ca' can use new objects defined in the
18659 * Add cool BIO that does syslog (or event log on NT).
18661 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18663 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18664 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18665 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18666 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18670 * Add preliminary config info for new extension code.
18674 * Make RSA_NO_PADDING really use no padding.
18676 *Ulf Moeller <ulf@fitug.de>*
18678 * Generate errors when private/public key check is done.
18682 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18683 for some CRL extensions and new objects added.
18687 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18688 key usage extension and fuller support for authority key id.
18692 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18693 padding method for RSA, which is recommended for new applications in PKCS
18694 #1 v2.0 (RFC 2437, October 1998).
18695 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18696 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18697 against Bleichbacher's attack on RSA.
18698 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18701 * Updates to the new SSL compression code
18703 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18705 * Fix so that the version number in the master secret, when passed
18706 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18707 (because the server will not accept higher), that the version number
18708 is 0x03,0x01, not 0x03,0x00
18710 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18712 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18713 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18714 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18718 * Support for RAW extensions where an arbitrary extension can be
18719 created by including its DER encoding. See `apps/openssl.cnf` for
18724 * Make sure latest Perl versions don't interpret some generated C array
18725 code as Perl array code in the crypto/err/err_genc.pl script.
18727 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18729 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18730 not many people have the assembler. Various Win32 compilation fixes and
18731 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18732 build instructions.
18736 * Modify configure script 'Configure' to automatically create crypto/date.h
18737 file under Win32 and also build pem.h from pem.org. New script
18738 util/mkfiles.pl to create the MINFO file on environments that can't do a
18739 'make files': perl util/mkfiles.pl >MINFO should work.
18743 * Major rework of DES function declarations, in the pursuit of correctness
18744 and purity. As a result, many evil casts evaporated, and some weirdness,
18745 too. You may find this causes warnings in your code. Zapping your evil
18746 casts will probably fix them. Mostly.
18750 * Fix for a typo in asn1.h. Bug fix to object creation script
18751 obj_dat.pl. It considered a zero in an object definition to mean
18752 "end of object": none of the objects in objects.h have any zeros
18753 so it wasn't spotted.
18755 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18757 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18758 Masking (CBCM). In the absence of test vectors, the best I have been able
18759 to do is check that the decrypt undoes the encrypt, so far. Send me test
18760 vectors if you have them.
18764 * Correct calculation of key length for export ciphers (too much space was
18765 allocated for null ciphers). This has not been tested!
18769 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18770 message is now correct (it understands "crypto" and "ssl" on its
18771 command line). There is also now an "update" option. This will update
18772 the util/ssleay.num and util/libeay.num files with any new functions.
18774 perl util/mkdef.pl crypto ssl update
18775 it will update them.
18779 * Overhauled the Perl interface:
18780 - ported BN stuff to OpenSSL's different BN library
18781 - made the perl/ source tree CVS-aware
18782 - renamed the package from SSLeay to OpenSSL (the files still contain
18783 their history because I've copied them in the repository)
18784 - removed obsolete files (the test scripts will be replaced
18785 by better Test::Harness variants in the future)
18787 *Ralf S. Engelschall*
18789 * First cut for a very conservative source tree cleanup:
18790 1. merge various obsolete readme texts into doc/ssleay.txt
18791 where we collect the old documents and readme texts.
18792 2. remove the first part of files where I'm already sure that we no
18793 longer need them because of three reasons: either they are just temporary
18794 files which were left by Eric or they are preserved original files where
18795 I've verified that the diff is also available in the CVS via "cvs diff
18796 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18797 the crypto/md/ stuff).
18799 *Ralf S. Engelschall*
18801 * More extension code. Incomplete support for subject and issuer alt
18802 name, issuer and authority key id. Change the i2v function parameters
18803 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18804 what that's for :-) Fix to ASN1 macro which messed up
18805 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18809 * Preliminary support for ENUMERATED type. This is largely copied from the
18814 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18816 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18818 * Make sure `make rehash` target really finds the `openssl` program.
18820 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18822 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18823 like to hear about it if this slows down other processors.
18827 * Add CygWin32 platform information to Configure script.
18829 *Alan Batie <batie@aahz.jf.intel.com>*
18831 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18833 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18835 * New program nseq to manipulate netscape certificate sequences
18839 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18844 * Fixes to BN code. Previously the default was to define BN_RECURSION
18845 but the BN code had some problems that would cause failures when
18846 doing certificate verification and some other functions.
18848 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18850 * Add ASN1 and PEM code to support netscape certificate sequences.
18854 * Add ASN1 and PEM code to support netscape certificate sequences.
18858 * Add several PKIX and private extended key usage OIDs.
18862 * Modify the 'ca' program to handle the new extension code. Modify
18863 openssl.cnf for new extension format, add comments.
18867 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18868 and add a sample to openssl.cnf so req -x509 now adds appropriate
18873 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18874 error code, add initial support to X509_print() and x509 application.
18878 * Takes a deep breath and start adding X509 V3 extension support code. Add
18879 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18880 stuff is currently isolated and isn't even compiled yet.
18884 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18885 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18886 Removed the versions check from X509 routines when loading extensions:
18887 this allows certain broken certificates that don't set the version
18888 properly to be processed.
18892 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18893 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18894 can still be regenerated with "make depend".
18898 * Spelling mistake in C version of CAST-128.
18900 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18902 * Changes to the error generation code. The perl script err-code.pl
18903 now reads in the old error codes and retains the old numbers, only
18904 adding new ones if necessary. It also only changes the .err files if new
18905 codes are added. The makefiles have been modified to only insert errors
18906 when needed (to avoid needlessly modifying header files). This is done
18907 by only inserting errors if the .err file is newer than the auto generated
18908 C file. To rebuild all the error codes from scratch (the old behaviour)
18909 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18910 or delete all the .err files.
18914 * CAST-128 was incorrectly implemented for short keys. The C version has
18915 been fixed, but is untested. The assembler versions are also fixed, but
18916 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18917 to regenerate it if needed.
18918 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18919 Hagino <itojun@kame.net>*
18921 * File was opened incorrectly in randfile.c.
18923 *Ulf Möller <ulf@fitug.de>*
18925 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18926 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18927 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18928 al: it's just almost always a UTCTime. Note this patch adds new error
18929 codes so do a "make errors" if there are problems.
18933 * Correct Linux 1 recognition in config.
18935 *Ulf Möller <ulf@fitug.de>*
18937 * Remove pointless MD5 hash when using DSA keys in ca.
18939 *Anonymous <nobody@replay.com>*
18941 * Generate an error if given an empty string as a cert directory. Also
18942 generate an error if handed NULL (previously returned 0 to indicate an
18943 error, but didn't set one).
18945 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18947 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18951 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18952 parameters. This was causing a warning which killed off the Win32 compile.
18956 * Remove C++ style comments from crypto/bn/bn_local.h.
18958 *Neil Costigan <neil.costigan@celocom.com>*
18960 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18961 based on a text string, looking up short and long names and finally
18962 "dot" format. The "dot" format stuff didn't work. Added new function
18963 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18964 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18965 OID is not part of the table.
18969 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18970 X509_LOOKUP_by_alias().
18974 * Sort openssl functions by name.
18978 * Get the `gendsa` command working and add it to the `list` command. Remove
18979 encryption from sample DSA keys (in case anyone is interested the password
18984 * Make *all* `*_free` functions accept a NULL pointer.
18986 *Frans Heymans <fheymans@isaserver.be>*
18988 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18991 *Anonymous <nobody@replay.com>*
18993 * s_server should send the CAfile as acceptable CAs, not its own cert.
18995 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18997 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18999 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19001 * Temp key "for export" tests were wrong in s3_srvr.c.
19003 *Anonymous <nobody@replay.com>*
19005 * Add prototype for temp key callback functions
19006 SSL_CTX_set_tmp_{rsa,dh}_callback().
19010 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19011 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19015 * X509_name_add_entry() freed the wrong thing after an error.
19017 *Arne Ansper <arne@ats.cyber.ee>*
19019 * rsa_eay.c would attempt to free a NULL context.
19021 *Arne Ansper <arne@ats.cyber.ee>*
19023 * BIO_s_socket() had a broken should_retry() on Windoze.
19025 *Arne Ansper <arne@ats.cyber.ee>*
19027 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19029 *Arne Ansper <arne@ats.cyber.ee>*
19031 * Make sure the already existing X509_STORE->depth variable is initialized
19032 in X509_STORE_new(), but document the fact that this variable is still
19033 unused in the certificate verification process.
19035 *Ralf S. Engelschall*
19037 * Fix the various library and `apps/` files to free up pkeys obtained from
19038 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19042 * Fix reference counting in X509_PUBKEY_get(). This makes
19043 demos/maurice/example2.c work, amongst others, probably.
19045 *Steve Henson and Ben Laurie*
19047 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19048 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19049 are no longer created. This way we have a single and consistent command
19050 line interface `openssl <command>`, similar to `cvs <command>`.
19052 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19054 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19055 BIT STRING wrapper always have zero unused bits.
19059 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19063 * Make the top-level INSTALL documentation easier to understand.
19067 * Makefiles updated to exit if an error occurs in a sub-directory
19068 make (including if user presses ^C) [Paul Sutton]
19070 * Make Montgomery context stuff explicit in RSA data structure.
19074 * Fix build order of pem and err to allow for generated pem.h.
19078 * Fix renumbering bug in X509_NAME_delete_entry().
19082 * Enhanced the err-ins.pl script so it makes the error library number
19083 global and can add a library name. This is needed for external ASN1 and
19084 other error libraries.
19088 * Fixed sk_insert which never worked properly.
19092 * Fix ASN1 macros so they can handle indefinite length constructed
19093 EXPLICIT tags. Some non standard certificates use these: they can now
19098 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19099 into a single doc/ssleay.txt bundle. This way the information is still
19100 preserved but no longer messes up this directory. Now it's new room for
19101 the new set of documentation files.
19103 *Ralf S. Engelschall*
19105 * SETs were incorrectly DER encoded. This was a major pain, because they
19106 shared code with SEQUENCEs, which aren't coded the same. This means that
19107 almost everything to do with SETs or SEQUENCEs has either changed name or
19108 number of arguments.
19110 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19112 * Fix test data to work with the above.
19116 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19117 was already fixed by Eric for 0.9.1 it seems.
19119 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19121 * Autodetect FreeBSD3.
19125 * Fix various bugs in Configure. This affects the following platforms:
19129 unixware-2.0-pentium
19134 * Eliminate generated files from CVS. Reorder tests to regenerate files
19135 before they are needed.
19139 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19143 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19145 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19146 changed SSLeay to OpenSSL in version strings.
19148 *Ralf S. Engelschall*
19150 * Some fixups to the top-level documents.
19154 * Fixed the nasty bug where rsaref.h was not found under compile-time
19155 because the symlink to include/ was missing.
19157 *Ralf S. Engelschall*
19159 * Incorporated the popular no-RSA/DSA-only patches
19160 which allow to compile a RSA-free SSLeay.
19162 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19164 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19165 when "ssleay" is still not found.
19167 *Ralf S. Engelschall*
19169 * Added more platforms to Configure: Cray T3E, HPUX 11,
19171 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19173 * Updated the README file.
19175 *Ralf S. Engelschall*
19177 * Added various .cvsignore files in the CVS repository subdirs
19178 to make a "cvs update" really silent.
19180 *Ralf S. Engelschall*
19182 * Recompiled the error-definition header files and added
19183 missing symbols to the Win32 linker tables.
19185 *Ralf S. Engelschall*
19187 * Cleaned up the top-level documents;
19188 o new files: CHANGES and LICENSE
19189 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19190 o merged COPYRIGHT into LICENSE
19191 o removed obsolete TODO file
19192 o renamed MICROSOFT to INSTALL.W32
19194 *Ralf S. Engelschall*
19196 * Removed dummy files from the 0.9.1b source tree:
19197 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19198 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19199 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19200 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19201 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19203 *Ralf S. Engelschall*
19205 * Added various platform portability fixes.
19209 * The Genesis of the OpenSSL rpject:
19210 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19211 Young and Tim J. Hudson created while they were working for C2Net until
19214 *The OpenSSL Project*
19216 ### Changes between 0.9.0b and 0.9.1b [not released]
19218 * Updated a few CA certificates under certs/
19222 * Changed some BIGNUM api stuff.
19226 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19227 DGUX x86, Linux Alpha, etc.
19231 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19232 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19237 * Add -strparse option to asn1pars program which parses nested
19240 *Dr Stephen Henson <shenson@bigfoot.com>*
19242 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19246 * DSA fix for "ca" program.
19250 * Added "-genkey" option to "dsaparam" program.
19254 * Added RIPE MD160 (rmd160) message digest.
19258 * Added -a (all) option to "ssleay version" command.
19262 * Added PLATFORM define which is the id given to Configure.
19266 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19270 * Extended the ASN.1 parser routines.
19274 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19278 * Added a BN_CTX to the BN library.
19282 * Fixed the weak key values in DES library
19286 * Changed API in EVP library for cipher aliases.
19290 * Added support for RC2/64bit cipher.
19294 * Converted the lhash library to the crypto/mem.c functions.
19298 * Added more recognized ASN.1 object ids.
19302 * Added more RSA padding checks for SSL/TLS.
19306 * Added BIO proxy/filter functionality.
19310 * Added extra_certs to SSL_CTX which can be used
19311 send extra CA certificates to the client in the CA cert chain sending
19312 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19316 * Now Fortezza is denied in the authentication phase because
19317 this is key exchange mechanism is not supported by SSLeay at all.
19321 * Additional PKCS1 checks.
19325 * Support the string "TLSv1" for all TLS v1 ciphers.
19329 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19330 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19334 * Fixed a few memory leaks.
19338 * Fixed various code and comment typos.
19342 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19343 bytes sent in the client random.
19345 *Edward Bishop <ebishop@spyglass.com>*
19349 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19350 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19351 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19352 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19353 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19354 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19355 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19356 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19357 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19358 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19359 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19360 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19361 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19362 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19363 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19364 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19365 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19366 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19367 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19368 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19369 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19370 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19371 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19372 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19373 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19374 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19375 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19376 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19377 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19378 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19379 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19380 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19381 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19382 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19383 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19384 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19385 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19386 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19387 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19388 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19389 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19390 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19391 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19392 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19393 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19394 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19395 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19396 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19397 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19398 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19399 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19400 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19401 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19402 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19403 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19404 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19405 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19406 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19407 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19408 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19409 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19410 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19411 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19412 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19413 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19414 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19415 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19416 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19417 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19418 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19419 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19420 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19421 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19422 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19423 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19424 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19425 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19426 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19427 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19428 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19429 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19430 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19431 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19432 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19433 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19434 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19435 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19436 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19437 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19438 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19439 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19440 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19441 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19442 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19443 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19444 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19445 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19446 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19447 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19448 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19449 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19450 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19451 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19452 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19453 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19454 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19455 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19456 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19457 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19458 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19459 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19460 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19461 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19462 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19463 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19464 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19465 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19466 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19467 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19468 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19469 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19470 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19471 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19472 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19473 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19474 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19475 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19476 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19477 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19478 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19479 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19480 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19481 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19482 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19483 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19484 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19485 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19486 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19487 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19488 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19489 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19490 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19491 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19492 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19493 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19494 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19495 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19496 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19497 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19498 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19499 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19500 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19501 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19502 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19503 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19504 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19505 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19506 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19507 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19508 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19509 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19510 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655